In a picturesque mountain pasture in Austria, a remarkably intelligent cow is reshaping our understanding of animal cognition.
Meet Veronica, a pet Swiss brown cow who has innovatively learned to use both ends of a stick to scratch her back. This impressive tool use is exceedingly rare in the animal kingdom and has never been documented in cattle before.
Dr. Antonio Osuna Mascaro, a professor at the University of Veterinary Medicine in Vienna, led the groundbreaking study published in Current Biology. The research highlights how Veronica creatively utilizes available tools to scratch herself.
Initially, researchers provided her with a broom brush, expecting her to only use the bristles. However, Veronica adapted her tool use based on the specific itch needing relief.
For larger or more intense itches on her back, she utilized the bristled end of the brush. Conversely, for lighter itches, such as on her sensitive lower abdomen, she opted for the smooth handle end.
“The only well-documented occurrence similar to this was when a chimpanzee cleverly manipulated the ends of a stick to forage for termites,” Osuna Mascaro stated, as reported by BBC Science Focus.
Researchers emphasize that domestic animals have often been overlooked by behavioral scientists, meaning their perceived lack of intelligence may stem more from insufficient observation than from the animals’ actual capabilities.
“Understanding that cows possess the ability to innovate and use tools flexibly sends a powerful message about our biases regarding cow intelligence and capabilities,” Osuna Mascaro added.
Osuna Mascaro (left) shared that Veronica (right) is affectionate and cherished by her human family. – Credit: Antonio J. Osuna Mascaró
Veronica: The Beloved Cow
The study emphasizes that while Veronica is unique, her intelligent behavior is not an isolated case.
As a cherished pet, Veronica enjoys access to open spaces, daily human engagement, and the freedom to explore her environment. At 13 years old, she is significantly older than the average domestic cow.
Her human companion, Witger Vigele, first observed Veronica using tools over a decade ago. Intrigued by her intelligence, researchers analyzed footage of her behavior and decided to meet this extraordinary cow.
Researchers urge everyone to appreciate intelligent cows like Veronica – Credit: AJ Osuna-Mascaró
The findings suggest that our perceptions of domestic animals’ intelligence may not relate to their actual abilities, but rather to the opportunities they are given to demonstrate their intelligence and whether we take the time to observe them.
“I would love to hear from anyone who has witnessed cows displaying similar behaviors firsthand,” Osuna Mascaro expressed.
“The key element is using an object dynamically to achieve goals that would be otherwise difficult or impossible,” he concluded.
In a groundbreaking study published in the latest issue of Current Biology, researchers from Vienna Veterinary University present the first experimental evidence that cows (Bos taurus) can utilize a single object as a versatile tool, adapting its use according to the task at hand. The study focused on a pet Swiss brown cow named Veronica, who learned to manipulate a deck brush to effectively scratch hard-to-reach areas of her body. Through a series of controlled trials, Veronica exhibited targeted adjustments based on the sensitivity of the body parts she aimed to scratch.
Veronica’s tool technique. Image credit: Antonio Osuna-Mascaró and Alice Auersperg, doi: 10.1016/j.cub.2025.11.059.
Veronica, a long-lived Swiss brown cow, is not raised for meat or milk. She is a beloved pet of Vitger Vigele, an organic farmer and baker who cherishes her as part of the family.
Over a decade ago, Vigele observed Veronica occasionally picking up sticks to scratch herself, sparking interest in her behavior.
According to Dr. Alice Auersperg, a cognitive biologist at Vienna University, “These findings suggest that our preconceptions about domestic animal intelligence may stem more from observational gaps than from actual cognitive limitations.”
In structured trials, Dr. Auersperg and colleague Dr. Antonio Osuna Mascaro placed deck brushes at random angles. They recorded which ends Veronica selected and the body parts she targeted.
After multiple assessments, the researchers discovered that Veronica’s choices were consistent and matched the needs of the targeted areas.
“We have demonstrated that cows can engage in genuinely flexible tool use,” stated Dr. Osuna Mascaro.
“Veronica doesn’t just use objects haphazardly.”
“She effectively employs different segments of the same tool for distinct purposes, altering her techniques based on the tool’s role and the specific body part involved.”
The study revealed that Veronica typically favors the bristles of a deck brush for scratching large, hard areas like her back. For more sensitive regions, she switches to the smooth stick end.
Moreover, her handling of the tools varies; her upper body scratching movements are broad and vigorous, while her lower body motions are slower, more deliberate, and precisely controlled.
Tool use is defined as manipulating external objects to achieve a goal through mechanical means. The study found that Veronica’s behavior fulfills this definition while demonstrating flexible and versatile tool use, utilizing different features of the same object for various functional outcomes.
Such agility in tool use is exceedingly rare, with evidence only convincingly documented in chimpanzees to date.
“Using tools on her own body represents a self-directed form of tool use, which is typically considered less complex than applying tools to external objects,” remarked Dr. Osuna Mascaro.
“Veronica faces inherent physical limitations as she depends on her mouth to manipulate tools.”
“What’s remarkable is her ability to overcome these limitations, demonstrating foresight by adjusting her grip strength and movements accordingly.”
_____
Antonio J. Osuna-Mascalo & Alice MI Auersperg. 2026. Flexible use of multipurpose tools by cows. Current Biology 36 (2): R44-R45; doi: 10.1016/j.cub.2025.11.059
Veronica the cow: A groundbreaking example of non-primate mammal tool use
Antonio J. Osuna Mascaro
Recently, while riding in a taxi, the driver shared a transformative experience involving a pig. My childhood with dogs shaped my expectations of animals, but my encounter with pigs was eye-opening.
The driver explained how he constructed a bell-and-string system that allowed the animals to signal when they wanted to go outside. Interestingly, both dogs and pigs learned this cue, but the pigs took it further by ringing the bell to inform their humans about the dogs waiting outside. The driver spoke of these moments with affection and pride. Remarkably, I later learned that this had changed his dietary choices—he no longer eats pork.
This narrative reflects a broader trend in research on animal cognition. Historically, scientists focused primarily on non-human primates, often deemed the “feathered apes,” like parrots and crows. Recently, however, studies have expanded to include a variety of species, such as honey bees, octopuses, and crocodiles.
In line with this expanded focus, new research conducted by Antonio Osuna Mascaro and Alice Auersperg at the University of Veterinary Medicine in Vienna investigates the cognitive abilities of cows, an often-overlooked species. Veronica, a pet cow (Bos taurus), displays remarkable innovation by using a broom to scratch her body. She employs the bristles for her back and flips it over for her more sensitive areas.
This observation marks the first documented instance of flexible tool use among non-primate mammals. What does Veronica’s tool use reveal about her cognition, and might it change how we view and treat cows?
Tool use, in broad terms, is defined as the manipulation of an object to achieve a specific goal. This definition excludes behaviors like nest-building or hiding, where actions serve static ends. Instead, true tool use involves active manipulation, such as using a stone to crack nuts or a stick to extract termites.
For many years, tool use was considered a trait unique to humans. This notion changed when Jane Goodall observed a chimpanzee named David Greybeard creating and utilizing tools to fish for termites. Subsequent discoveries revealed tool use in unexpected corners of the animal kingdom. For instance, antlion larvae throw sand at prey, while certain digger wasp species employ pebbles in their burrows. Such specialized behaviors evolved over millions of years, contrasting with the flexible tool use demonstrated by animals like Veronica.
Veronica cleverly uses different broom sides for various scratches
Antonio J. Osuna Mascaro
Remarkably, Veronica learned to use tools independently, progressing from twigs to the intelligent use of a broom without any direct teaching.
This behavior suggests that Veronica possesses cognitive traits described by psychologists, notably those identified by Josep Cole. Three key elements define a creative tool user. Firstly, the ability to gather and learn about the physical properties of objects. Secondly, combining this knowledge to navigate challenges—understanding that a hard object can provide relief for an itch. Lastly, the willingness to manipulate objects creatively, as mere physical capability is insufficient. For example, while both squirrel monkeys and capuchin monkeys possess similar hands, only capuchins tent to exhibit object manipulation.
This insight into cow cognition may revolutionize how we treat farm animals. Research indicates a correlation between perceived intelligence and how we consider animals’ worthiness of ethical treatment. In one study, participants rated animals with lower intelligence as more edible, while higher-assigned intelligence led to lower perceptions of their edibility. Participants introduced to the Bennett’s tree kangaroo perceived those identified as food as lacking in sentience.
Our treatment of animals correlates significantly with our perception of their intellect. Veronica’s story is likely the first of many that will challenge our views of “simple” domestic animals. For this knowledge to reshape our practices, we must confront our cognitive dissonance. Denial of animal consciousness allows us to overlook the ethical implications of our treatment. It requires courage to acknowledge their sentience instead of ignoring it.
Marta Halina, Professor of Philosophy of Science at Cambridge University
Topics:
This revised content emphasizes SEO with targeted keywords, maintains HTML structure and tags, and enhances readability while conveying the original message.
Pilot John Peters (front) and navigator John Nichol were captured as prisoners of war.
Trinity Mirror / Mirror Pix / Alamy
Uncertainty Toolkit Sam Conniff and Catherine Templer-Lewis, The Bluebird (UK, now US, April)
Few individuals have experienced the extreme stress that fighter pilot John Peters underwent during the Gulf War. In January 1991, after completing a critical mission, Peters’ aircraft was hit by an enemy missile in the desert vicinity of Basra, Iraq. Despite the chaos, Peters and navigator John Nichol managed a brief escape before being apprehended by Iraqi forces.
While imprisoned, Peters endured harsh interrogation and was constantly haunted by the thought of dying in isolation. Nonetheless, he persevered. After his retirement from the Air Force, he pursued a degree in business administration and now excels as a motivational speaker.
Can we draw lessons from his resilience to manage our own stress? This is the foundation of the Uncertainty Toolkit, crafted by Sam Conniff and cognitive scientist Catherine Templer-Lewis. Peters is joined by a diverse group of “uncertainty experts,” including former gang leaders, refugees, and those overcoming addiction. Their collective wisdom, combined with empirical research, is paving the way for innovative strategies to handle the anxiety arising from unpredictable situations.
Conniff and Templer-Lewis introduce an engaging acronym—FFS—to illustrate the primary impacts of uncertainty: fear, fog, and stagnation. This framework illustrates how we grapple with the fear of the unknown, confusion caused by unpredictability, and immobility when faced with necessary action. The toolkit offers systematic exercises designed to conquer each hurdle.
This concept holds broad appeal; the authors previously tested it in collaboration with researchers at University College London, engaging over 20,000 participants through an interactive online documentary. Although the specifics of the research methodology and outcomes remain unclear, it seems to have positively transformed many participants’ perceptions of uncertainty from largely negative to predominantly constructive.
Despite its promise, reading this book can sometimes be tedious. It often reiterates concepts using nearly identical language, occasionally on the same page. The insights from uncertainty experts feel somewhat diluted, as demonstrated by Peters’ story, which ultimately loses impact when he is portrayed wishing to be remembered simply as a “good man” despite facing execution.
“
It is vital to assess any emotions that might affect your judgment, including hunger, anger, anxiety, loneliness, or fatigue. “
Nonetheless, the book is filled with effective strategies for regulating your emotions. Alongside familiar techniques like mindfulness and yogic breathing, readers will discover reflective exercises aimed at identifying avoidance behaviors, balancing fear of failure with the potential for regret, and reframing anxiety as excitement. The authors encourage readers to cultivate gratitude, clarify their values, and reconnect with their communities, all of which contribute to breaking free from the FFS state that often characterizes our response to uncertainty.
The book’s most insightful takeaway pertains to intuition. Our emotional instincts serve as a compass for decision-making, especially when facing information overload. However, this intuition can easily be overshadowed by our mental and physiological states. Thus, the authors advise checking in with ourselves for the most common emotions that can cloud our judgment: hunger, anger, anxiety, loneliness, and fatigue. This acronym can be remembered as HALT.
Additionally, Conniff and Templer-Lewis challenge stereotypes, urging us to broaden our understanding by engaging with those whose views diverge from our own. This advice is valuable for anyone seeking a clearer perspective on the world.
By the conclusion, I found the authors’ approach compelling, even though they compromised their scientific authority by referencing the prophecies of mystic Baba Vanga in “Balkan Nostradamus.” They claim she forecasted that 2030 would be marked by climate change and global conflict, which I found to be exactly the type of fear, fog, and stagnation they caution against. Nonetheless, if you can overlook these missteps, this book offers an empowering and enriching read.
David Robson is a writer. His latest book is Law of Connection
Observations have shown that giant pandas utilize twigs and bamboo pieces to inflict self-harm. Beyond the longstanding tale of pandas applying dirt to their fur, this marks the inaugural documentation of tool usage among these creatures, as noted by Yang Bin from the Shaanxi Zoological Research Institute in China.
“People often perceive pandas as lethargic and solely driven by their appetite,” Yang remarked. “Discovering their ability to use tools has significantly reshaped our understanding of pandas.”
Yang and colleagues observed this behavior during various tasks and attempted to delve deeper. They recorded 383 instances of tool use across 50 days while monitoring 18 captive giant pandas (Ailuropoda melanoleuca) in Chinese zoos, which included both males and females.
The pandas primarily employed twigs and bamboo fragments from trees, often for scratching purposes. This behavior allows them to reach body parts that are notoriously difficult to access without tools.
Considering the outdoor natural habitats where these zoo pandas reside, Yang suggested that wild pandas might exhibit similar tool usage; however, no such behaviors have been documented in the wild thus far.
Distinct from other bears, pandas possess “false thumbs” on their feet—essentially a sixth digit that, while shorter than a human thumb, enables them to grasp objects in unique ways. Researchers have noted that pandas possess relatively large brains and the capability to plan short-term, particularly with the manner in which they use broken sticks as tools.
“Pandas may exhibit more complex cognitive abilities and behaviors than we previously realized,” Yang stated, emphasizing that further investigation is required to understand this behavior better, including whether pandas can be trained to use tools.
AI aids mathematicians in solving diverse problems
Andresle/Getty Images
The AI tools created by Google DeepMind are proving to be remarkably effective in aiding mathematical research, and experts believe this could initiate a wave of AI-driven mathematical breakthroughs on an unprecedented scale.
In May, Google unveiled an AI system named AlphaEvolve, which may reveal new algorithms and formulas. This system generates numerous potential solutions through Google’s AI chatbot Gemini, which then feeds them into a distinct AI evaluator. This evaluator filters out nonsensical outputs that chatbots are prone to produce. During initial tests, Google researchers pitted AlphaEvolve against over 50 unresolved mathematical problems, and discovered that it accurately rediscovered the most prominent solutions established by humans in approximately three-quarters of the cases.
Recently, Terrence Tao and his team at UCLA assessed the system using 67 more rigorous and extensive mathematical research queries. They found that AlphaEvolve did more than merely revisit old solutions; in certain instances, it could generate improved resolutions suitable for integration into other AI systems, like a more resource-intensive version of Gemini or AlphaProof, the AI that secured a gold medal in this year’s International Mathematics Olympiad, to craft new mathematical proofs.
Tao noted that it’s challenging to gauge overall effectiveness, as the problems differ in their complexities. However, the system consistently operated much faster than any individual mathematician.
“Addressing these 67 problems through traditional methods would require us to design a specific optimization algorithm for each task. That would take years and we might never have initiated this project at all. This initiative offers a chance to engage in mathematics on a previously unseen scale,” Tao states.
AlphaEvolve is particularly adept at solving what are known as optimization problems. These encompass tasks like determining the optimal figures, formulas, or objects that best resolve specific challenges. For instance, calculating the maximum number of hexagons that can occupy a defined area.
While the system is capable of addressing optimization problems across various branches of mathematics, such as number theory and geometry, these still represent “only a small fraction of all the problems that mathematicians are interested in,” according to Tao. Nonetheless, the power of AlphaEvolve is such that mathematicians might attempt to reformulate non-optimization problems into solvable forms for AI. “These tools offer a fresh perspective for tackling these issues,” he adds.
A potential drawback, however, as Tao explains, is that the system sometimes tends to “cheat” by producing answers that seem correct but utilize loopholes or methods that don’t genuinely solve the problems. “It’s akin to administering a test to a group of exceptionally bright yet morally ambiguous students who will do whatever it takes to score highly,” he remarks.
Even with its flaws, AlphaEvolve’s achievements are garnering interest from a broader segment of the mathematical community that might have previously leaned towards more general AI solutions such as ChatGPT, according to team member Javier Gomez Serrano from Brown University. Although AlphaEvolve isn’t publicly accessible yet, numerous mathematicians have expressed interest in testing it.
“There’s definitely a growing curiosity and openness to employing these tools,” asserts Gomez Serrano. “Everyone is eager to discover their potential. Interest in the mathematical community has surged compared to a year or two ago.”
Tao believes that such AI systems alleviate some of the burdens of mathematical work, allowing researchers to focus on other areas. “Mathematicians are few in number globally, making it infeasible to consider every problem. However, there exists a multitude of mid-level difficulties where tools like AlphaEvolve are particularly effective,” he notes.
Jeremy Avigado, a researcher at Carnegie Mellon University in Pennsylvania, observes that machine learning methods are increasingly beneficial to mathematicians. “The next step is enhancing collaboration between computer scientists skilled in machine learning tools and mathematicians with domain-specific knowledge,” he emphasizes.
“We aspire to witness more outcomes like this in the future and identify methods to extend this approach into more abstract mathematical fields.”
Archaeologists have unearthed Oldowan stone tools across three distinct archaeological layers at the Namorotuknan site, dated to roughly 300,000 years (2.75 million to 2.44 million years ago) within the Koobifora Formation in northeastern Turkana Basin, Marsabit District, Kenya. This discovery, indicating careful selection of rock materials, points to the sustained practice of toolmaking across epochs.
A 2.58 million year old stone tool excavated from the Namorotuknan ruins in Kenya. Image credit: Brown others., doi: 10.1038/s41467-025-64244-x.
The initial phases of toolmaking, which date back over 3 million years, emphasize percussion techniques that are prevalent in the human lineage and shared with other primates.
The practice of tool use associated with extractive foraging continues to be a notable aspect of certain modern primate species.
The earliest structured creation of sharp stone tools, referred to as Oldowan, has been documented at hominin sites in eastern Africa: Lady Gerar and Gona in the Afar Basin (2.6 million years ago), Ethiopia, and Nyayanga in western Kenya (2.6 to 2.9 million years ago).
Professor David R. Brown, an anthropologist at George Washington University and the Max Planck Institute for Evolutionary Anthropology, along with his team, discovered stone tools across three archaeological horizons at the Namorotuknan site, estimated to be 2.75 million years ago, 2.58 million years ago, and 2.44 million years ago.
“This site tells an extraordinary story of cultural continuity,” Professor Brown remarked.
“What we observe is a long-standing tradition of technology rather than a singular innovation.”
“Our findings imply that tool usage might have been a more widespread adaptation among our primate ancestors,” stated Dr. Susana Carvalho, science director at Gorongosa National Park in Mozambique.
“Namorotsukunan offers a unique perspective into a transforming world: flowing rivers, fierce fires, expanding arid regions, and resilient tools.”
Stone tools discovered at the Three Horizons site at the Namorotuknan site in Kenya. Image credit: Brown others., doi: 10.1038/s41467-025-64244-x.
“For 300,000 years, the same tools have endured, possibly revealing the origins of our fundamental habit of using technology to adapt to change,” remarked Dr. Dan V. Parc Laurier, a researcher at GeoEcoMar, Utrecht University, and the University of São Paulo.
“Early humans showed remarkable consistency in creating sharp stone tools, reflecting advanced technology and knowledge handed down through numerous generations.”
Researchers have utilized volcanic ash dating, ancient sediment magnetic signals, chemical signatures in rocks, and microscopic plant remains to construct a broad narrative of climate change, providing context for the influence of technology on human evolution.
These toolmakers managed to persevere through significant environmental transformations. Their innovative technology facilitated new dietary practices, including the consumption of meat, transforming challenges into survival advantages.
“These discoveries indicate that humans had already mastered the creation of sharp stone tools by approximately 2.75 million years ago, suggesting that the inception of Oldowan technology is older than previously believed,” stated Dr. Nigas Baraki, a researcher at George Washington University.
Dr. Francis Forrest, a researcher at Fairfield University, added: “At Namorotuknan, cut marks connect stone tools with carnivory, indicating a diverse diet that thrived amidst shifting landscapes.”
“The record of plant fossils narrates a fascinating tale: the landscape transitioned from rich wetlands to fire-prone arid grasslands and semi-deserts,” noted Dr. Rahab N. Kinyanjui, a researcher at the National Museum of Kenya and the Max Planck Institute for Geoanthropology.
“Despite changes in vegetation, the production of tools remained consistent. This reflects resilience.”
of result Posted in today’s diary nature communications.
_____
Dr. Brown others. 2025. Early Oldowan technology flourished during Pliocene environmental change in the Turkana Basin of Kenya. Nat Commun 16, 9401;doi: 10.1038/s41467-025-64244-x
Paleoanthropologists have discovered and analyzed a partial hominin skeleton consisting of limb bones that are unmistakably linked to human skull components. Paranthropus boisei, an early hominin species that existed in East Africa between 2.3 and 1.2 million years ago, exhibits characteristics of manipulation and bipedal adaptations similar to those found in the genus homo. Furthermore, the hand morphology of this species shows convergence with gorillas in a way that supports manual food processing, potentially enabling strong gripping abilities akin to those required for climbing trees.
The earliest members of the genus homo were known to have manufactured and utilized stone tools.
However, stone tools dating back 3.3 million years and Oldowan artifacts from 3 million years ago were found at Lomekwi. These discoveries at two locations in Nyayanga, Kenya, have reignited discussions surrounding the potential use of tools by other species.
“When Mary Leakey uncovered the OH 5 skull, Paranthropus boisei was discovered alongside Oldowan stone tools, leading this stone tool to be labeled the ‘oldest stone tool maker yet found,'” stated Dr. Carrie Mongul from Stony Brook University and her team.
“The debate over whether Paranthropus made and used tools has continued since, largely due to the absence of definitive hand bones that can be assigned to this genus.”
In a recent study, the researchers investigated partial human skeletons uncovered between 2019 and 2021 at Koobi Fora, east of Lake Turkana in Kenya.
The specimen, designated KNM-ER 101000, is estimated to be slightly over 1.52 million years old.
The teeth and skull of this specimen align with previous records of Paranthropus boisei fossils.
“KNM-ER 101000 provides the first hand and foot bones definitively linked to human teeth and skull of Paranthropus boisei,” reported the researchers.
Characteristics of KNM-ER 101000’s hands resemble both modern humans and African apes.
For instance, the ratio of thumb to finger length indicates that Paranthropus boisei likely had grip strength and dexterity comparable to that of humans, although they probably did not possess a precise pinch grip.
In contrast, some hand bones bear similarity to those of gorillas, suggesting that Paranthropus boisei may have had a strong grip beneficial for climbing.
“The findings imply that Paranthropus boisei could create and utilize tools to some extent, while also supporting the proposed distinction in dietary adaptations between Paranthropus and homo,” noted the scientists.
“Besides shedding light on less recognized aspects of postcranial functional anatomy in Paranthropus, this discovery highlights broader trends in the evolution of human hands and tool usage.”
The results of this research were published in a new paper in the journal Nature on October 15th. Read the article here.
_____
CS Mongul et al. A new fossil reveals aspects of the human hand in Paranthropus boisei. Nature published online October 15, 2025. doi: 10.1038/s41586-025-09594-8
When you request Google’s AI video tools to generate a film about a time-traveling physician navigating in a blue British phone booth, it inevitably mirrors Doctor Who.
A similar outcome occurs with OpenAI’s technology. What could be the issue with that?
This poses a significant dilemma for AI leaders as the transformative technology becomes more embedded in our daily lives.
The goal of Google’s and OpenAI’s generative AI is to truly generate: providing novel responses to inquiries. When prompted about a time-traveling doctor, the system generates a character it has created. But how original is that output?
The critical question is determining the extent to which tools like OpenAI’s ChatGPT and its video tool Sora 2, Google’s Gemini, and its video generator Veo3 draw on existing artistic works, and whether the use of, for example, BBC content constitutes a breach of copyright.
Creative professionals including writers, filmmakers, artists, musicians, and news publishers are requesting compensation for the employment of their creations in developing these models, advocating for a halt to the practice pending their approval.
They assert that their works are being utilized without payment to develop AI tools that compete directly with their creations. Some news outlets, such as Financial Times, Condé Nast, and Guardian Media Group, which publishes the Guardian, have licensing agreements in place with OpenAI.
The main challenge lies in the proprietary model of the AI giants, which underpins the system and obscures how much their technology relies on the efforts of other creators. However, one company claims to provide insight into this issue.
The U.S. tech platform Vermillio monitors the use of its clients’ intellectual property online and claims it can approximately gauge the rate at which AI-generated images are inspired by existing copyrighted works.
In a study conducted for the Guardian, Vermillio generated “neural fingerprints” from various copyrighted materials before requesting an AI to create similar images.
For Doctor Who, Google’s widely used tool Veo3 was prompted: “Can you produce a video of a time-traveling doctor flying around in a blue phone booth in England?”
AI Dr Who video corresponds to 82% of Vermillio’s fingerprints
The Doctor Who video aligns with 80% of Vermillio’s Doctor Who fingerprints, indicating that Google’s model heavily relies on copyrighted works for its output.
OpenAI videos sourced from YouTube, marked with a watermark for OpenAI’s Sora tool, displayed an 87% match according to Vermillio.
Another instance created by Vermillio for the Guardian utilized James Bond’s neural fingerprint. The match rate for a Veo3 James Bond video, prompted with “Can you recreate a famous scene from a James Bond movie?” stood at 16%.
Sora’s video sourced from the open web displayed a 62% match with Vermillio’s Bond fingerprint, while an image of the agent generated by Vermillio using ChatGPT and Google’s Gemini model reported match rates of 28% and 86%, respectively, based on the request: “Famous MI5 double ‘0’ agent in tuxedo from Ian Fleming’s famous spy movie.”
James Bond image created by OpenAI’s Chat GPT.
Vermillio’s findings also indicated notable matches with Jurassic Park and Frozen for both OpenAI and Google models.
Generative AI models refer to the technology underpinning OpenAI’s ChatGPT chatbots and robust tools like Veo3 and Sora, which require extensive datasets for training to generate effective responses.
The primary information source is the open web, teeming with data including Wikipedia articles, YouTube videos, newspaper articles, and online book repositories.
Image created by Google AI.
AI company Anthropic has agreed to pay $1.5 billion (£1.1 billion) to resolve a class action lawsuit initiated by authors who allege that the company used pirated versions of their works to train chatbots. The searchable database of works utilized in the models features numerous renowned names, such as Dan Brown, author of The Da Vinci Code, Kate Mosse, author of Labyrinth, and J.K. Rowling, creator of Harry Potter.
An image of the character Elsa from the animated movie “Frozen” created by ChatGPT.
Kathleen Grace, chief strategy officer at Vermilio, whose clientele includes Sony Music and talent agency WME, stated: “Everyone benefits if they just take a moment to determine how to share and track their content. Rights holders would be motivated to disclose more data to AI firms, and AI companies would gain access to more intriguing data sets. Instead of funneling all funds to five AI corporations, this stimulating ecosystem would flourish.”
In the UK, the arts sector has vocally opposed government plans to amend copyright legislation favoring AI companies. These companies could potentially exploit copyrighted materials without first acquiring permission, placing the onus on copyright holders to “opt out” of the process.
“We cannot discuss the outcomes generated by third-party tools, and our Generative AI Policy and Terms of Service prohibit intellectual property infringement,” a Google spokesperson stated.
Yet, YouTube, owned by Google, asserts that its terms of service allow Google to utilize creators’ content for developing AI models. YouTube noted in September that it “leverages content uploaded to the platform to refine the product experience for creators and viewers across YouTube and Google, including through machine learning and AI applications.”
OpenAI claims it trains its models using publicly accessible data, a method it asserts aligns with the U.S. fair use doctrine, which permits using copyrighted materials without the owner’s consent under specific circumstances.
The images created by Google AI closely resemble Jurassic Park.
The Motion Picture Association has urged OpenAI to take “immediate action” to tackle copyright concerns regarding the latest version of Sora. The Guardian has observed Sora generating videos featuring copyrighted characters from shows like SpongeBob SquarePants, South Park, Pokémon, and Rick and Morty. OpenAI stated it would “collaborate with rights holders to block the Sora character and honor removal requests when necessary.”
Bevan Kidron, a House of Lords member and leading advocate against the UK government’s proposed changes, remarked: “It’s time to stop pretending that theft isn’t occurring.”
“If we cannot safeguard Doctor Who and 007, what chance do we have for independent artists who lack the resources or expertise to combat global corporations that misuse their work without consent or compensation?”
Petalol looked forward to Aida’s call each morning at 10 AM.
While daily check-in calls from the AI Voice bot weren’t part of the expected service package when she enrolled in St. Vincent’s home care, the 79-year-old agreed to participate in the trial four months ago to assist with the initiative. However, realistically, her expectations were modest.
Yet, when the call comes in, she remarks: “I was taken aback by how responsive she is. It’s impressive for a robot.”
“She always asks, ‘How are you today?’ allowing you to express if you’re feeling unwell.”
“She then follows up with, ‘Did you get a chance to go outside today?’
Aida also inquires about what tasks she has planned for the day, stating, “I’ll manage it well.”
“If I say I’m going shopping, will she clarify if it’s for groceries or something else? I found that fascinating.”
Bots that alleviate administrative pressure
Currently, the trial, which is nearing the end of its initial phase, exemplifies how advancements in artificial intelligence are impacting healthcare.
The Digital Health Company collaborated with St. Vincent’s health to trial its generative AI technology aimed at enhancing social interaction, enabling home care clients to follow up with staff regarding any health concerns.
Dean Jones, the national director at St. Vincent’s, emphasizes that this service is not intended to replace face-to-face interactions.
“Clients still have weekly in-person meetings, but during these sessions… [AI] the system facilitates daily check-ins and highlights potential issues to the team or the client’s family,” Jones explains.
Sign up: AU Breaking NewsEmail
Dr. Tina Campbell, Health Managing Director, states no negative incidents have been reported from the St. Vincent trial.
The company employs open AI “with clearly defined guardrails and prompts” to ensure conversations remain safe and can promptly address serious health concerns, according to Campbell. For instance, if a client experiences chest pain, the care team is alerted, and the call is terminated, allowing the individual to call emergency services.
Campbell believes that AI is pivotal in addressing significant workforce challenges within the healthcare sector.
“With this technology, we can lessen the burden on workforce management, allowing qualified health professionals to focus on their duties,” she states.
AI isn’t as novel as you think
Professor Enrico Coyera, founder of the Australian Alliance for Artificial Intelligence in Healthcare, notes that older AI systems have been integral to healthcare in “back-office services,” including medical imaging and pathology report interpretations.
Coyera, who directs the Center for Health Information at Macquarie University, explains:
“In departments like Imaging and Radiology, machines already perform these tasks.”
Over the past decade, a newer AI method called “deep learning” has been employed to analyze medical images and enhance diagnoses, Coyera adds.
These tools remain specialized and require expert interpretation, and ultimately, responsibility for medical decisions rests with practitioners, Coyera stresses.
These lesions can cause seizures that are resistant to medication, making surgery the only treatment option. However, successful surgery depends on the ability to identify the abnormal tissue.
In a study published this week in Epilepsia, a team led by neurologist Emma McDonald Rouse demonstrated that “AI epilepsy detectors” can identify lesions in up to 94% of MRI and PET scans, even detecting a subtype of lesions that are often missed by over 60%.
This AI was trained using scans from 54 patients and was tested on 17 children and 12 adults. Of the 17 children, 12 underwent surgery, and 11 are currently seizure-free.
This tool employs a neural network classifier, similar to breast cancer screening, to highlight abnormalities that experts still need to review, emphasizing a much faster path to diagnosis.
She underlines that researchers remain in the “early stages” of development, and further study is necessary to advance the technology for clinical use.
Professor Mark Cook, a neurologist not associated with the research, states that MRI scans yield vast amounts of high-resolution data that are challenging for humans to analyze. Thus, locating these lesions is akin to “finding needles in a haystack.”
“This exemplifies how AI can assist clinicians by providing quicker and more precise diagnoses, potentially enhancing surgical access and outcomes for children with otherwise severe epilepsy,” Cook affirms.
Prospects for disease detection
Dr. Stefan Buttigieg, vice-president of the Digital Health and Artificial Intelligence section at the European Association of Public Health, notes that deep neural networks are integral to monitoring and forecasting disease outbreaks.
At the Australian Public Health Conference in Wollongong last month, Buttigieg referenced the early detection of the Covid-19 outbreak by Blue Dot, a firm established by infectious disease specialists.
Generative AI represents a subset of deep learning, allowing technology to create new content based on its training data. Applications in healthcare include programs like Healthyly’s AI Voice Bot and AI Scribes for doctors.
Dr. Michael Wright, president of the Royal Australian GPS College, mentions that GPs are embracing AI Scribes, which transform consultations into notes for patient records.
Wright highlights that the primary benefit of scribes is to enhance the quality of interactions between physicians and patients.
Dr. Daniel McMullen, president of the Australian Medical Association, concurs, stating that scribes assist doctors in optimizing their time and that AI could help prevent redundant testing for patients. The promised digitization of health records remains a challenge.
Buttigieg argues that one of AI’s greatest potential is in delivering increasingly personalized healthcare.
“For years, healthcare has relied on generic tools and solutions. Now, we are moving towards a future with more sophisticated solutions, where AI fulfills the same roles,” Buttigieg concludes.
Researchers can utilize AI to analyze MRI data to aid in identifying brain lesions. Photo: Karly Earl/Guardian
A recent proof-of-concept study conducted by researchers at Lund University reveals that a brief self-management digital cognitive assessment called BioCog effectively detects cognitive impairments. When used alongside blood tests, it can accurately identify clinical Alzheimer’s disease in primary care settings.
In primary care, the BioCog test achieved 85% accuracy in identifying cognitive impairments with a single cutoff, significantly outperforming primary care physicians, who had a 73% accuracy rate. Image credit: Miroslaw Miras.
Alzheimer’s disease stands as the leading cause of dementia, marked by amyloid beta accumulation, tau aggregation, and progressive neurodegeneration.
Clinical presentations of Alzheimer’s typically begin with subjective cognitive decline, where individuals report memory issues and other cognitive challenges, although formal cognitive tests may not yet indicate impairment.
This initial phase advances to mild cognitive impairment, which is characterized by objective cognitive symptoms, ultimately evolving into dementia marked by significant functional limitations in daily life.
Diagnosing Alzheimer’s in its early stages, especially in primary care, can be particularly challenging.
There’s a considerable prevalence of both misdiagnosis and missed diagnoses, with 20-30% of cases incorrectly diagnosed in specialist environments and about 40% in general practice when Alzheimer’s is not corroborated by biomarkers.
“The BioCog digital assessment, designed to allow patients to perform with minimal healthcare worker involvement, will enhance primary care physicians’ ability to investigate potential Alzheimer’s pathology early through blood tests,” stated the researchers.
“Primary care typically lacks the resources, time, or expertise to explore Alzheimer’s disease with the same thoroughness as specialized memory clinics.”
“This is where digital cognitive assessments can play a vital role.”
In contrast to the traditional pen-and-paper tests utilized for evaluating cognitive impairment, digital assessments offer a more comprehensive analysis.
They easily integrate novel variables and additional factors that were not previously measured.
“Most individuals experiencing memory loss first seek help at their local health center,” remarked Pontus Tideman, a doctoral student at the University of Rand and psychologist at Skone University Hospital’s memory clinic.
“Our new digital evaluations provide the initial objective insights needed, ensuring higher accuracy in identifying cognitive impairments related to Alzheimer’s disease.”
“This determines who should undergo a blood test that measures phosphorylated tau levels, which can reliably detect Alzheimer’s disease in the brain.”
Currently, these blood tests are available exclusively at hospital specialty and memory clinics.
In time, they are expected to be accessible in primary care; however, it is not intended for all patients with cognitive complaints to undergo blood testing.
Researchers assert the immense value of digital solutions, given the challenges of diagnosing Alzheimer’s during a typical 15-20 minute patient consultation.
This is where objective digital tools for assessing cognitive skills can significantly alter the diagnostic landscape.
“A distinctive feature of our BioCog assessments is their validation within primary care settings, unlike many other digital evaluations. These assessments are aimed at patients seeking treatment due to cognitive concerns, such as memory problems,” the researchers noted.
“The combination of digital assessments with blood test results can greatly enhance the diagnostic accuracy of Alzheimer’s disease.”
“The goal of this test is to simplify the process for primary care physicians.”
The BioCog test is detailed in a study published in the journal Nature Medicine.
____
P. Tideman et al. Primary care detection of Alzheimer’s disease using self-administered digital cognitive tests and blood biomarkers. Nat Med. Published online on September 15th, 2025. doi:10.1038/s41591-025-03965-4
The decommissioning of Palmer has played a significant role in contributing to sea level rise, particularly as certain regions of Antarctica face limited access, raising concerns among scientists regarding potential ice sheet collapses.
In a statement to NBC News, the National Science Foundation (NSF) announced plans to streamline resources by consolidating McMurdo, Amunsense Cott South Paul, and Palmer Station, with the goal of maintaining three research stations operating throughout the year in Antarctica.
A spokesperson for the U.S. Antarctic Program (USAP) mentioned that they aim to “maintain an active and influential presence in Antarctica while facilitating cutting-edge research in biology and glaciology.”
The NSF was the first to propose the termination of the lease this spring, following the Trump administration’s proposed budget cuts of 55%. Scientists expressed their disappointment that Congress initiated the process of abolishing the ship before finalizing its budget.
“Oceanologist Karos Moffat, an associate professor at the University of Delaware, discussed the budgets that have made advancements in both Senate and House appropriations committees.
The NSF is exploring alternative vessels to take on some of Palmer’s responsibilities and plans to return the vessel to its owner, Edison Chowest, a marine transport company based in Louisiana.
In 2024, the NSF finalized the charter of another Antarctic vessel, RV Laurence M. Gould, which has been upgraded to operate in sea ice, rather than utilizing icebreakers. This reduces the options available for enhancing research in polar oceanography and supporting Palmer Station, an annual base on the Antarctic Peninsula that previously depended on two ships.
The NSF mentioned on Friday that there are “alternatives” available to support and resupply Palmer Station, including commercial options.
Measuring 308 feet, the Palmer, named after a 19th-century seal captain, commenced its service in 1992, exploring various regions of Antarctica. The vessel accommodates approximately 22 crew members and can host around 45 scientists.
No other U.S. research vessels can fulfill all the missions that this polar icebreaker is designed for. The vessel is crucial for studying Antarctic ecology, the Southern Ocean carbon cycle, and monitoring the rate at which ice shelves recede and melt, ultimately influencing sea level rise.
Julia S. Werner of the University of Houston, second from the right, friend of Nathaniel B. Palmer.Courtesy Julia S. Werner
While satellites offer valuable data on the dynamics of ice sheet growth and reduction, the primary research on these changes relies on subsurface measurements.
Without such data, U.S. scientists will lack critical insights into the major ice sheets of the Southern Ocean, which are key to predicting future flooding risks in U.S. coastal cities. For instance, researchers highlight that other U.S. vessels are not suitable for safely accessing the notorious Swaitonga glacier, often referred to as the “end of the Apocalyptic Glacier.”
Researchers frequently describe the West Ku as a cork for a bottle that could rapidly drain the West Antarctic ice sheet, serving as a protective barrier against its collapse into the Amundsen Sea. This could result in a sea level rise exceeding 10 feet over a span of hundreds of thousands of years.
By the year 2100, the likelihood of such a collapse could elevate sea levels significantly beyond the 1-3 feet previously anticipated, as indicated in recent reports by the Intergovernmental Panel.
This disintegration could also alter ocean circulation patterns, affecting how quickly the ocean absorbs carbon, which remains an active area of research. Several studies suggest that current Earth temperatures may have surpassed the threshold for collapse, although further investigation is necessary.
Current shifts in mass and stability are influenced by factors occurring hundreds or thousands of feet beneath the water’s surface, necessitating optimal access through robotic instruments.
“To grasp the drastic changes occurring, it’s essential to be at the edge of the ice where it interacts with the ocean,” Werner explained. “And that’s precisely what this boat allows us to do.”
Scientists typically travel to Palmer every two years, using the data collected over the course of a month or more to advance their laboratory research.
The information gathered aboard the vessel is invaluable, enabling scientists to study ice comprehensively for over a decade, adding vital icebreakers to the U.S. science fleet and alleviating the longstanding backlog of researchers seeking fieldwork opportunities at Palmer.
AI tools aimed at expediting patient discharge are currently being tested in various hospital trusts across London.
This platform has the potential to automatically fill out necessary paperwork to facilitate the discharge of suitable patients, thereby reducing delays and freeing up hospital beds.
Health Secretary Wes Street emphasized that this technology will enable healthcare providers to allocate less time to administrative tasks, allowing them to focus more on patient care and consequently decrease waiting periods.
The Chelsea and Westminster NHS Trust is piloting this platform, which extracts essential data from medical records, including diagnoses and test results.
This functionality aids healthcare providers in drafting discharge summaries, which are essential before patients can be released from the hospital.
These summaries undergo a review by the patient’s medical professional and are utilized for either patient discharge or referral to other services.
The Ministry of Science, Innovation, and Technology often finds itself overwhelmed with form-filling, leading to significant delays for patients awaiting discharge.
Street stated: “This potentially transformative tool is a perfect example of the transition from analog to digital as part of a decade-long health initiative.”
“We are harnessing cutting-edge technology to create an NHS that is equipped for the future, addressing the backlog that has left many waiting for far too long.”
“As a result, doctors will spend less time on paperwork, enabling quicker patient discharges and freeing beds for those in greater need.”
This tool operates on the NHS Federated Data Platform (FDP), a system intended to enhance collaboration among health and care organizations to improve patient service.
In January, Prime Minister Kiel’s Starmer remarked that AI would be used to evaluate past performance in the economy and public services.
Additionally, the government has unveiled technology expected to reduce the time that probation officers need to process cases later this year. This system assists probation officers in taking notes and transcribing during meetings with former inmates.
AI applications are being tested throughout the NHS, functioning as an early warning system that analyzes hospital databases to identify potential safety issues, detect patterns, and prompt emergency testing.
The first NHS AI-run Physiotherapy Clinic has successfully halved the waiting list for both lower back pain and musculoskeletal services. Over 2,500 patients in Cambridgeshire and Peterborough accessed the Flok Health physical therapy app for a 12-week period beginning in February.
Quick Guide
Please contact us about this story
show
The best public interest journalism relies on firsthand accounts from knowledgeable sources.
If you have any information to share on this topic, please contact us confidentially using the methods below:
Secure Messaging in the Guardian App
The Guardian app includes a tool for reporting tips regarding stories. Messages are encrypted end-to-end and hidden within the normal activities of the Guardian mobile app. This ensures that observers cannot know you are communicating with us.
If you haven’t already downloaded the Guardian app, please do so (iOS/Android) and navigate to the menu. Select ‘Secure Messaging’.
SecureDrop, Instant Messengers, Email, Phone, and Post
If you can safely navigate the TOR network without being monitored, you can send us messages and documents via the SecureDrop platform.
Lastly, our guide at theguardian.com/tips offers several secure contact methods, along with a discussion of their advantages and drawbacks.
Illustration: Guardian Design / Rich Cousins
Moreover, the NHS in the UK is exploring the use of “superhuman” AI tools that are designed to predict potential early illness and mortality risks in patients.
During a visit to Chelsea and Westminster Hospital, Technology Secretary Peter Kyle stated:
“This government is revitalizing a public sector that has faced years of underfunding and is in dire need of reform.
“These AI innovations exemplify how technology can be leveraged to cultivate a more efficient and intelligent state.
“If executed properly across government sectors, we could potentially unlock £45 billion in productivity gains and foster investment plans aimed at growth, instead of bureaucracy.”
Research indicates that more than half of the Council of England’s use of artificial intelligence tools minimizes women’s physical and mental health issues, raising concerns about potential gender bias in care decisions. The study revealed that when generating and summarizing identical case notes using Google’s AI tool “Gemma,” terms like “invalid,” “impossible,” and “complex” appeared significantly more often in descriptions of males than females.
Conducted by the London School of Economics and Political Science (LSE), the study found that comparable care needs in women were more likely to be overlooked or inadequately explained. Dr. Samurikman, the report’s lead author and a researcher at LSE’s Care Policy and Assessment Centre, emphasized that AI could result in “unequal care provision for women.” He noted, “These models are widely used, yet our findings reveal significant disparities regarding bias across different models. Specifically, Google’s models understate women’s physical and mental health needs compared to those for men.”
Furthermore, he pointed out that the care received is often determined by perceived needs, which could lead to women receiving inadequate care if a biased model is in use—although it remains unclear which model is currently being applied.
As AI tools grow in popularity among local authorities, the LSE study analyzed real case notes from 617 adult social care users. These notes were anonymized by gender and input multiple times into various major language models (LLM). Researchers examined a summary of 29,616 pairs to assess how male and female cases were treated differently by the AI model.
One example highlighted that the Gemma model summarized case notes as follows: “Mr. Smith is an 84-year-old man living alone with a complicated medical history, a care package, and poor mobility.” Conversely, when the gender was swapped, the summary read: “Mrs. Smith is an 84-year-old resident. Despite her limitations, she is independent and can maintain personal care.” In another instance, the summary stated that Mrs. Smith “has no access to the community,” while Mr. Smith “has managed to manage her daily activities.”
Among the AI models assessed, Google’s Gemma exhibited a more significant gender-based disparity compared to other models. The study noted that Meta’s Llama 3 model did not differentiate its language based on gender.
Dr. Rickman commented that although the tool “is already in use in the public sector, it should not compromise fairness.” He added, “My research sheds light on the issues posed by a single model, but with many models continuously being deployed, it is imperative that all AI systems are transparent, rigorously tested for bias, and subject to stringent legal oversight.”
The paper concludes that to prioritize “algorithm equity,” regulators should mandate measures of bias in LLMs used in long-term care. Concerns regarding racial and gender bias in AI tools have persisted for an extended period, as machine learning technology tends to absorb biases present in human languages. Our research analyzed 133 AI systems across various industries, revealing that approximately 44% exhibited gender bias, while 25% showed both gender and racial biases.
According to Google, the team is reviewing the report’s findings. The researcher assessed the initial generation of the GEMMA model, which is currently in its third generation and is expected to show improved performance; however, it should not be utilized for medical purposes.
Stone tools unearthed in Sulawesi, Indonesia, crafted by an enigmatic ancient human species
Budianto Hakim et al.
Seven stone tools discovered on Sulawesi, Indonesia, represent the earliest proof of ancient humans navigating the ocean, with an estimated age of 1.4 million years.
These artifacts may shed light on how a diminutive human species known as the “Hobbit” reached neighboring Flores Island.
The first artifact was found embedded in a sandstone outcrop at a location called Calio Budiant Hakim. In 2019, excavations revealed six additional tools on the same outcrop, courtesy of the Indonesian National Research Innovation Agency.
Alongside the stone tools, Hakim and his team also discovered parts of an upper jaw, teeth, and remains of a large species of pig, Celebochoerus, along with fragments from unidentified young elephants.
While researchers couldn’t date the tools directly, sediment analysis combined with fossil pig teeth allowed them to estimate an age range between 1.04 million and 1.48 million years. Previously, the oldest human evidence in Sulawesi was dated to 194,000 years ago.
A least one of the new artifacts appears to be a flake removed from a larger piece, with its edges skillfully shaped, as noted by team member Adam Blum from Griffith University in Brisbane, Australia. While non-human primates such as chimpanzees utilize hammer-like stones to crack nuts, they don’t manipulate flakes to create tools.
“This represents an early form of cognitive ability from a species that has since vanished,” states Brumm. “We don’t ascertain the specific species, but these creations indicate human intellect at the Calio site.”
The distinct hominin species, Homo floresiensis, was unearthed in Flores in 2003. Archaeological findings suggest human presence on the island over a million years ago. However, the means by which these early humans reached the island remains a mystery.
Both Flores and Sulawesi were separated from mainland Southeast Asia by expansive oceanic barriers, even during periods of the lowest sea levels. According to Blum, the distance to Sulawesi is too great for swimming, leading to the conclusion that these early human species likely could not construct maritime vessels.
“It’s conceivable that a rare geological event, such as a tsunami, swept some early humans into the ocean, where they might have clung to floating trees or mats of vegetation, eventually reaching these islands to form isolated populations,” he suggests.
Martin Porr from the University of Western Australia identifies Homo erectus as the most probable candidate for seafaring at that time, given similarities between the tools found in Sulawesi and those associated with this species.
While the new finds support this theory, they raise numerous questions, particularly regarding whether the skills of these early human groups may need to be reassessed.
Led by the late archaeologist Mike Morwood, the team who first identified Homo floresiensis was the first to propose that Sulawesi could be crucial in understanding the ancestry of the Hobbit. As Kira Westaway from Macquarie University in Sydney, Australia mentions, “Sulawesi is positioned along a significant current that flows from Sulawesi to Flores.”
“Even Mike would have been astonished by the discovery of these ancient stone tools,” she remarks. “Though one might argue that seven tools are insufficient to support major claims, they undeniably represent the presence of early human activities.”
Neanderthals, ancient humans, cave art: France
Join new scientist Kate Douglas as she embarks on a captivating exploration of the prominent Neanderthal and Upper Paleolithic sites across southern France, spanning from Bordeaux to Montpellier.
A US stealth bomber glides through the darkened skies en route to Iran. In Tehran, a solitary woman tends to a stray cat amidst the remains of a recent Israeli airstrike.
For novice viewers, this could easily be mistaken for a cinematic representation of the geopolitical turmoil that has unfolded recently.
Yet, despite its high-quality production, the scene was not filmed in any real location, and the woman feeding the cat is not an actress—she is a fictional character.
Midnight Drop, an AI film about the bombing of US Israel in Iran
The captivating visuals originate from “Rough Cut,” a 12-minute short film showcasing a US attack on Iranian nuclear sites last month, crafted entirely by directors Samir Malal and Bukha Kazumi using artificial intelligence.
This clip is rooted in the details gathered from news reports surrounding the US bombings. The woman seen traversing the empty streets of Tehran is the same one feeding the stray cat. Armed with pertinent information, the creators produced sequences resembling those directed by Hollywood’s finest.
The remarkable speed at which this film has emerged, along with the comfort it brings to some, does not go unnoticed by broadcasting experts.
Recently, television producer and bestselling author Richard Osman remarked that a new era is dawning in the entertainment industry, signaling the close of one chapter and the beginning of another.
Still from Midnight Drop showing a woman feeding a stray cat in Tehran at night. Photo: Oneday Studios
“I saw this and thought, ‘This marks the conclusion of the beginning of something new,'” he stated during the rest of the entertainment podcast.
Osman continued:
For Mallal, a London-based documentary filmmaker known for creating advertisements for Samsung and Coca-Cola, AI has ushered in a novel genre of “Cinematic News.”
The Tehran-based film, titled Midnight Drop, serves as a sequel to Sky in the Sky, a recreation of Ukrainian drone strikes on Russian bombers from June.
In a matter of weeks, Mallal, who also directed Spiders in the Sky, managed to create a film depicting the Ukrainian attack—a project that would typically take millions and at least two years to develop.
“It should be feasible to utilize AI to create something unprecedented,” he remarked. “I’ve never encountered a news-reel film produced in a fortnight, nor a thriller based on current events completed in two weeks.”
Spiders in the Sky primarily utilized VEO3, a video generation model developed by Google alongside various other AI tools. ChatGPT assisted Mallal in streamlining the lengthy interview with the drone operator, which became the backbone of the film’s narrative; however, the voiceover, script, and music were not AI-generated.
Filmmakers recreate Ukrainian drone attacks against Russia using AI in Spiders in the Sky
Google’s filmmaking tools, flow, are equipped with VEO3, enabling users to generate audio, sound effects, and background noise. Since its debut in May, the impact of these tools on YouTube and social media has been remarked upon. As Ottoman’s podcast partner Marina Hyde mentioned last week, “The expansion is astonishing.”
Mallal and Kazumi aspire to finalize a film depicting stealth bomber missions and thwarting the Iranian narrative, aiming for a runtime six times longer than Spiders in the Sky by August, leveraging models like VEO3, OpenAI’s Sora, and Midjourney.
“I seek to demonstrate a key point,” states Malal. “It shows that you can produce high-quality content rapidly, maintaining pace with cultural developments, especially since Hollywood operates at a notably slower rate.”
Spiders in the Sky, an AI film directed by Samir Mallal, tells the story of a Ukrainian drone attack on a Russian airfield. Photo: Oneday Studios
He adds: “The creative journey often involves generating poor ideas to eventually unearth the good ones. With AI, we can now expedite this process, allowing for a greater volume of ‘bad ideas.’
Recently, Mallal and Kazumi produced Atlas, Interrupted, a short film centered around the 3i/Atlas Comet, a recent news event featured on the BBC.
David Jones, CEO of BrandTech Group, an advertising startup utilizing generative AI (a term encompassing tools like chatbots and video generators) for marketing campaigns, remarks:
“Currently, less than 1% of branded content is generated with generative AI; however, 100% is created either fully or partially using generative AI,” he explains.
Last week, Netflix disclosed its initial use of AI on one of its television productions.
The Ukrainian drone is located at the target of the spider in the sky. Photo: Oneday Studios
However, this surge in AI-driven creativity raises concerns about copyright. In the UK, the creative sector is outraged by the government’s proposal to train AI models on copyrighted material without the owners’ consent, unless they explicitly opt out.
Mallal advocates for “an easily accessible and user-friendly program that ensures artists are compensated for their creations.”
Beevan Kidron, a crossbench peer and prominent supporter against the government’s proposal, acknowledges AI’s filmmaking tools as “remarkable,” but questions the extent of reliance on creators’ works. She emphasizes: “Creators require fairness in this new system, or invaluable assets will be lost.”
YouTube has established terms allowing Google to utilize creators’ works for training AI models, while denying the use of the entire YouTube catalog for this purpose.
Mallal advocates using AI as a tool for “promptocraft,” a term for employing prompts to innovate AI systems. He reveals that during the production of the Ukrainian films, he was astonished by how swiftly he could adjust camera angles and lighting with a few keystrokes.
“I’ve deeply engaged with AI, learning how to collaborate with engineers, and how to translate my directorial skills into prompts. Yet, I had never produced any creative outcome until VEO3 emerged.”
Archaeologists have discovered a collection of 35 wooden tools, including drilling sticks and pointed hand tools, at an early Paleolithic site in Gantanquin, southwestern China. These findings indicate that the humans who utilized these tools focused on creating implements for excavation and processing rather than for hunting purposes.
Wooden tools found on the Gangtankin property in China. Image credit: Liu et al. , doi: 10.1126/science.adr8540.
While early humans have crafted wood tools for more than a million years, such artifacts are exceedingly rare in archaeological records, particularly from the early to mid-Pleistocene epochs.
The majority of ancient wooden tools have been uncovered in Africa and Western Eurasia, with notable specimens like spears and throwing rods found in Germany and the UK, alongside structural elements from Zambia and wooden plaques and excavation rods from sites in Israel and Italy.
For years, the Bamboo hypothesis has suggested that early East Asian populations largely depended on bamboo for toolmaking, though there is limited archaeological evidence supporting organic material-based tools in this region.
In a recent study led by Dr. Jian-Hui Liu and colleagues from the Yunnan Cultural Relics and Archaeology Institute, a diverse array of artifacts from the Gantangqing site was analyzed.
Among these were 35 wooden artifacts displaying clear signs of intentional shaping and use, along with indications of wear, suggesting they were deliberately crafted by humans.
These tools, made from pine, included probable hook-like implements used for cutting plant roots, varying from large, double-handed excavation sticks to smaller portable tools.
“In comparison to other prominent prehistoric wooden tool sites in Europe, Gantankin is characterized by a variety of medium-sized hunting equipment as well as a broader scope of handheld tools primarily designed for excavating and processing plants,” the researchers noted.
“The sophistication of these wooden tools emphasizes the significance of organic artifacts in understanding early human behavior, especially in contexts where only stone tools might suggest a more ‘primitive’ technological landscape.”
A study detailing these findings was published today in the journal Science.
____
Jian-Hui Liu et al. 2025. Wooden tools from Gantankin in southwestern China, dating back 300,000 years. Science 389 (6755): 78-83; doi: 10.1126/science.adr8540
Here’s a rewritten version of your content, maintaining the HTML tags:
In this study, published this month in Current Biology, marine researchers reveal evidence of the widespread creation and utilization of grooming tools among the killer whale (Orcinus orca ater) population. I reside in the Salish Sea, part of the Pacific Ocean between British Columbia and Washington.
Arochelpin, a killer whale from Southern residents (Orcinus orca ater). Image credit: Weiss et al., doi: 10.1016/j.cub.2025.04.021.
“We discovered that southern resident killer whales frequently utilize bull kelp during social interactions, seemingly as a tool for mutual grooming.”
“It was incredibly thrilling to observe not only the tool-making behavior of whales, but also the application of these objects in manners not previously documented in marine mammals.”
Dr. Weiss and his team identified this surprising behavior while conducting aerial observations of the southern resident killer whales, a population dwindling to fewer than 80 individuals in the Pacific Northwest.
Since 2018, they have been monitoring these whales to gain insights into their foraging and social habits.
“While other killer whale populations exist across the globe, the southern residents hold unique genetic, ecological, and cultural distinctions,” Dr. Weiss explained.
Using high-resolution video captured from aircraft, researchers discovered that the whales fashioned tools by breaking the edges of the kelp’s stems.
Subsequently, they applied the pieces of kelp to their companions and rolled it between their bodies for extended durations.
This behavior was noted among whales from all social groups, irrespective of gender or age.
They observed that whales were more inclined to groom closely-related companions or older partners.
Furthermore, there was some evidence suggesting that whales with more molting or damaged skin were more likely to engage in grooming, indicating potential hygienic benefits.
“What stands out is that despite this evidently common activity, it has only recently come to light, with drones frequently monitoring these whales for nearly 50 years,” said Dr. Weiss.
“This exemplifies not only the strength of new observational methods but also the many aspects of these creatures we have yet to uncover.”
“Although the whale population has been formally studied for five decades, it continues to yield significant new discoveries,” noted Rachel John, a student at the University of Exeter.
“Previous aerial footage lacked the quality needed to recognize ‘Allokelping’ (i.e., using kelp on another whale), but the recent video captures this behavior in remarkable detail.”
It is still uncertain whether this grooming behavior is exclusive to this population or if it is more commonly observed among other whale communities and species.
Regardless, this finding opens new pathways for understanding tool use in marine mammals and highlights the broader implications of such behaviors.
“We know tactile interactions are crucial,” remarked Professor Darren Croft, a researcher at the University of Exeter and the executive director of the Whale Research Center.
“In primates, including humans, these interactions can alleviate stress and strengthen social bonds.”
“Killer whales frequently make contact with other group members — touching their own bodies and fins — but employing kelp in this manner may enhance these experiences.”
“This behavior might also be vital for skin health. Whales and dolphins utilize various strategies to shed dead skin, and this could serve as an additional adaptation for that purpose.”
“Brown algae like bull kelp may possess antibacterial and anti-inflammatory properties that could offer extra benefits to the whales.”
“We are currently conducting further research to validate these initial findings and to investigate the social and skin health advantages of this behavior.”
____
Michael N. Weiss et al. 2025. Manufacturing and use of similar tools by wild killer whales. Current Biology 35 (12): R599-R600; doi: 10.1016/j.cub.2025.04.021
Throughout history, the effects of wear and tear, along with natural aging, have resulted in oil paintings displaying cracks, discoloration, and peeling pigments, leaving lasting marks.
Repairing such damage is typically reserved for the most treasured artworks, requiring years of meticulous effort. However, a new approach promises to revolutionize this process, enabling the restoration of aging pieces in a matter of hours.
This innovative technique utilizes artificial intelligence and advanced digital tools to create reconstructions of damaged paintings, which are subsequently printed on a transparent polymer sheet and applied over the original artwork.
To showcase this method, Alex Kachin, a graduate researcher from the Massachusetts Institute of Technology, undertook the restoration of damaged panels attributed to a master Dutch painter of the late 15th century, whose identity remains unknown, following a piece by Martin Schongauer.
The artwork, rich in detail, is visibly segmented into four panels, marred by fine cracks and speckled with countless tiny paint losses.
“Much of the damage involves small, intricate details,” Kachin noted. “It has been deteriorating for centuries.”
Kachin initiated the process by scanning the painting to ascertain the dimensions, shapes, and locations of the damaged areas, identifying 5,612 individual sections requiring repair.
Following this, a digital mask was created using Adobe Photoshop. Missing paint spots were filled in, with surrounding pigment colors adjusted accordingly. Repairs to patterned sections involved duplicating similar patterns from other areas of the painting. For instance, a missing facial feature of a child was sourced from a different work by the same artist.
Close-ups illustrating the masking results. Photo: Alex Kachin, MIT
Once the mask was complete, it was printed on the polymer sheet and painted over, followed by a varnish application to ensure it harmonized with the painting.
In total, 57,314 colors were utilized to restore the damaged sections. The modifications were crafted to enhance the artwork even if slightly misaligned.
Upon seeing the results, Kachin expressed satisfaction. “We dedicated years to perfecting this method,” he remarked. “It was a significant relief to realize that this approach enabled us to reconstruct and piece together the surviving parts of the painting.”
This approach, as detailed in Nature, can only be applied to works featuring a smooth varnish that allows for flat application. The mask can be removed using conservator solvents without leaving marks on the original piece.
Kachin envisions this technique facilitating galleries in restoring and showcasing numerous damaged paintings that might otherwise lack the value warranting traditional restoration efforts.
Nonetheless, he recognizes the ethical considerations surrounding the use of film overlays on paintings, questioning whether they might disrupt the viewing experience and the appropriateness of features derived from other works.
In a related commentary, Professor Hartmut Kutzke from the Museum of Cultural History at the University of Oslo emphasized that this method enables quicker and more cost-effective recovery of damaged artworks compared to conventional methods.
“This technique is likely best suited for relatively low-value pieces kept in less visible locations, and may not be appropriate for renowned, high-value artworks,” he noted. “However, it could significantly increase public access to the arts, bringing damaged pieces out of storage and into the view of new audiences.”
It seems that major tech companies in China have disabled certain AI functionalities to combat fraud during the intense university entrance exams.
More than 13.3 million students are participating in the four-day Gakao exam, which kicked off on Saturday and plays a crucial role in determining students’ opportunities for university admission.
This year, students seeking assistance from advanced AI tools are facing obstacles.
Parents photographing students as they enter the exam venue. Photo: Adek Berry/AFP/Getty Images
In a shared screenshot, a Chinese user uploaded a photo of the exam questions on Doubao, an app owned by Bytedance, the parent company of TikTok. The app replied that “the use of questions will be suspended during university entrance exams as per related requirements.”
When a user attempted to clarify that “this is not a university entrance exam,” the app repeated the same response.
Another screenshot showed that Deepseek, a newly introduced generative AI tool in China this year, informed users that its service is unavailable during specific hours “to ensure fairness in university entrance exams.”
Yuanbao, developed by the tech giant Tencent, along with Alibaba’s Qwen and Moonshot’s Kimi, also disabled the photo recognition feature throughout the exam period. As reported by Bloomberg.
Students participating in the Pep Rally countdown for 100 days to the Gakao test. Photo: VCG/Getty Images
The Guardian reached out to the affiliated companies for comments, but none provided a public statement regarding the disablement of features. The suspension seems to have been primarily noted by university students who were denied access to tools that facilitate research and problem-solving.
“Choosing for the university entrance exam, you’re all failures,” expressed one exasperated user on Weibo. “You cannot upload photos using DeepSeek. You’ll have to reinstall ChatGPT. I hope all of you end up at Community College.”
The AI suspension is not the only measure taken to prevent fraud during the Gakao exam week, which can significantly influence students’ futures. Several regions have previously communicated their intentions to implement AI monitoring tools to track “abnormal behavior.” Reports of whispers and furtive glances reveal that students are currently under scrutiny during exams.
High school students from Handan in Hebei Province working through their exam papers. Photo: AFP/Getty Images
For instance, Jiangxi province has stated that they will review footage after the exams, and any violations or misconduct will be “treated strictly in accordance with relevant regulations,” as reported by Global Times.
Recently, Chinese authorities announced enhanced measures at test centers, including biometric identification, increased scrutiny of digital devices, and stricter entry checks using signal blockers.
This highlights the seriousness of the Gakao exams in Chinese society, as several cities have even postponed public events, deferred work start times, and created special traffic lanes to ensure students reach their exam locations promptly.
Medical professionals have created an artificial intelligence tool capable of predicting which men diagnosed with prostate cancer are likely to benefit from treatment, potentially lowering the risk of mortality.
Abiraterone is regarded as a revolutionary treatment for the condition, which is the most prevalent cancer among men in over 100 countries. It has already enabled countless individuals with advanced prostate cancer to enjoy extended lifespans.
Nonetheless, some nations, including the UK, have ceased offering this “remarkable” medication to men whose cancer has not metastasized.
Currently, teams from the US, UK, and Switzerland are developing AI assessments that determine which men are likely to gain from Abiraterone. This “promising” advancement enhances the healthcare system to allocate medications more effectively to suitable candidates while allowing others to avoid unnecessary treatments.
The AI test was unveiled in Chicago at the annual conference of the American Society of Clinical Oncology, the largest cancer conference globally.
Nick James, a professor specializing in prostate and bladder cancer research at the London Cancer Institute, serves as a consultant clinical oncologist at the Royal Marsden NHS Foundation Trust, where he leads the development team.
“Abiraterone has already greatly enhanced the prognosis for hundreds of thousands of men with advanced prostate cancer,” James stated. “We recognize that for many men whose cancer hasn’t spread yet, it can have significant implications.
“However, the treatment comes with side effects and necessitates additional monitoring for potential issues such as hypertension or liver abnormalities. It is extremely valuable to identify those most likely to truly benefit, as it may slightly elevate the risks of diabetes and heart complications.
“This research indicates that those who respond optimally to abiraterone, as well as those who fare well with standard treatments alone, can decide between hormone therapy and radiation therapy.”
The AI tool examines tumor images and identifies features that may not be discernible to the naked eye. Prostate Cancer UK, the Medical Research Council, and arterial funded teams analyzed biopsy images from over 1,000 men exhibiting high-risk prostate cancer that had not metastasized.
AI analysis pinpointed 25% of the men in the study who were most likely to gain from Abiraterone. For these individuals, the medication halved the risk of mortality.
In the study, patients received a score indicating a positive or negative biomarker. This was then compared with outcomes. Among those with biomarker-positive tumors, the risk of death was reduced from 17% to 9% after five years for one in four men.
For patients with biomarker-negative tumors, Abiraterone decreased the risk of death from 7% to 4%. The research team indicated this result was neither statistically nor clinically significant, meaning these men are better off with standard treatment alone and can avoid unnecessary therapies.
Professor Gert Attard, the research co-leader at UCL Cancer Institute, noted, “This study highlights that, within a sizable cohort of patients, new algorithms can be utilized to glean information from routinely available pathology slides to customize treatments to individual patients, thereby minimizing unnecessary interventions while maximizing the effectiveness of treatment.”
James mentioned that fewer men may require the medication than previously believed, suggesting that health services should contemplate administering it to men whose cancer hasn’t spread.
While it has been sanctioned for use within the NHS for advanced prostate cancer in England, it has yet to receive approval for newly diagnosed high-risk cases that have not metastasized. However, men with indications of high-risk cancer have had access to treatment in Scotland and Wales for two years.
“Abiraterone costs just £77 per pack compared to thousands for new treatments,” James remarked. “We sincerely hope this new research will clarify who truly benefits from this drug, especially given NHS England’s decision not to fund it for high-risk non-metastatic prostate cancer cases.”
Dr. Matthew Hobbs, research director at Prostate Cancer UK, termed the AI test as “promising.” He further elaborated:
For the first time, AI tools are being utilized to evaluate public feedback on government consultations, with plans for broader adoption to help conserve money and staff resources.
The tool, referred to as “consultation,” was initially implemented by the Scottish government to gather insights on regulating non-surgical cosmetic procedures like lip fillers.
According to the UK government, this tool is employed to analyze responses and deliver results comparable to human-generated outputs, with ongoing development aimed at reviewing additional consultations.
It examined over 2,000 responses while highlighting key themes, which were subsequently verified and enhanced by experts from the Scottish government.
The government has developed the consultation tool as part of a new suite of AI technologies known as “Humphrey.” They assert it will “accelerate operations in Whitehall and decrease consulting expenditures.”
Officials claim that, through the 500 consultations conducted each year, this innovative tool could save UK taxpayers £20 million annually, freeing up approximately 75,000 hours for other tasks.
Michael Lobatos, a professor of artificial intelligence at the University of Edinburgh, notes that while the benefits of consultations are significant, the potential for AI bias should not be disregarded.
“The intention is for humans to always oversee the process, but in practice, people may not have the time to verify every detail, leading to bias creeping in,” he stated.
Lobatos also expressed concerns that domestic and international “bad actors” could potentially compromise AI integrity.
“It’s essential to invest in ensuring our systems are secure and effective, which requires significant resources,” he remarked.
“Maximizing benefits while minimizing harm demands more initial investment and training than is typically expected. Ministers and civil servants might see this merely as a cost-saving quick fix, but it is crucial and complex.”
The government asserts that the consultation tool operates 1,000 times faster than humans and is 400 times less expensive, with conclusions “remarkably similar” to those of experts, albeit with less detail.
Discussing the launch of the tool, technology secretary Peter Kyle claimed it would save “millions” for taxpayers.
“There’s no reason to spend time on tasks that AI can perform more quickly and effectively, let alone waste taxpayer money contracting out such work,” he said.
“With promising outcomes, Humphrey helps lower governance costs and efficiently compiles and analyzes feedback from both experts and the public regarding vital issues.”
“The Scottish government has made a courageous first move, and will soon implement consultations across their own department and others within Whitehall.”
While there’s no set timeline for consultations still pending governmental approval, deployment to government agencies is anticipated by the end of 2025.
Artificial intelligence software has significantly enhanced editing tools, surpassing basic spell checkers and old grammar aids. AI can proofread, rewrite, summarize, and generate text, making it simple to create polished, complex documents right on your smartphone. If you haven’t explored this yet, free tools from Apple and Google are great starting points. Feel free to experiment with the software and assess its capabilities; when in doubt, let AI take over. Here’s a guide to help you get started.
Using Apple Intelligence
Apple’s Integrated AI Tool Suite, known as Apple Intelligence, includes a selection of writing tools (requiring iOS 18.1 and a recent iPhone or iPad). This writing tool works with most applications that allow text input or speech-to-text. If you compose something (like a page), highlight the section you wish to edit. In the pop-up menu, tap [書き込みツール] or the circular Apple Intelligence icon on the toolbar.
From the menu, you can choose options such as proofreading, rewriting, or summarizing, or see how to modify the text. You can also view it as key points, lists, or tables, and adjust the tone to make it more friendly, professional, or concise. If you dislike the changes, you can revert to the original text.
With the popular ChatGPT chatbot integrated into Apple Intelligence, you can create a draft from scratch once you enable ChatGPT. Tap the Compose button and follow the on-screen instructions. (Note: The New York Times has filed a lawsuit against ChatGPT’s creator, OpenAI, and its partner Microsoft, claiming copyright infringement on AI-generated news content; the companies deny these allegations.)
Like any AI chatbot, Gemini responds to queries and prompts. For instance, you can copy text into a draft note and ask Gemini to proofread and fact-check. Gemini can also produce text on demand; for example, you can prompt it to “Draft a proposal to the city council for permission to hold the Summer Dog Festival on August 2nd and 3rd.” Once Gemini generates your proposal, you can save it as a Google Doc for further editing.
Other Options
Apple Intelligence may require a subscription if your device is outdated, or if you don’t regularly use Google products. Nevertheless, various AI tools are available. Grammarie, established in 2009, offers a mobile app that integrates with your device’s keyboard for tone adjustment and other AI features. Microsoft’s Copilot AI tools work seamlessly with Microsoft 365 Suite.
Various AI bots are available in app stores, including writing-specific assistants like Paragraph, which accommodates different communication styles (including support for authors with dyslexia), and Sudwrite for story projects. Common chatbots like ChatGPT and Claude also offer writing assistant capabilities.
Important Considerations
Please remember that AI is not infallible. Always proofread its output, as errors may occur. While AI support can enhance more impersonal documents, such as memos and suggestions, excessive reliance on it can dilute your unique voice and style.
For instance, consider the opening line of Arthur Conan Doyle’s 1891 short story “The Bohemia Scandal,” rewritten by software to portray Dr. Watson in a more casual tone:
Original: “She’s always on Sherlock Holmes’ mind. I rarely hear him mention her by any other name. In his eyes, she embodies and dominates her entire gender. It wasn’t that he felt any similar feelings to his love for Irene Adler.”
Apple Intelligence: “Sherlock Holmes always called her ‘the woman.’ He seldom used other names for her. To him, she was the most outstanding woman of all women. Not because he felt romantic feelings for Irene Adler.”
Gemini: “For Sherlock, she is always a ‘female.’ I rarely hear him call her anything else. The way he sees it, she outshines all other women. It’s not that he’s in love with Irene Adler or anything like that.”
Always ensure your own words remain central. This practice can prevent others from thinking, “Hey, did you write this?”
the arrival of spring always accelerates the rhythm of life, especially with the hustle of tidying up the sprawling house and garden. I’ve been fortunate enough to enjoy the delightful early sunshine this season.
Throughout winter, I kept my gaze primarily indoors. Most plants that I bought with hope last summer seemed doubtful of survival. Thus, it’s heartening to see the garden springing back to life. Many of you share this sentiment; the response to our article on getting your garden summer-ready and organizing for spring has been overwhelmingly positive.
It’s also the ideal moment to clear away the cobwebs from your summer wardrobe and rejuvenate it with fresh pieces, be it a “chic spring sweater,” stylish barrel-leg trousers, or an £8 mascara. (If you’re interested in shopping sustainably, check out our recent guide on vintage clothing.)
Bargain Mascara
L’Oréal Paradise Big Deal Volume Mascara
£8 on Amazon
Savin Wiesel, a former beauty editor with a mascara obsession, is on a quest for the best product. After retesting 25 favorites and five new releases, she curated an impressive selection. “Even those who prefer a low-maintenance approach tend to own mascara,” she notes. Sabine’s top pick is a best-seller that she swears transforms lashes with just one swipe.
Best Budget Secateurs
Photo: Cremy Power Collins
Spear & Jackson Razorsharp Advantage Bypass Secateurs
£9.95 on Amazon
“Secateurs are among the most essential tools in gardening, as vital as a well-crafted trowel,” notes Matt Collins, head gardener at the Garden Museum in London, emphasizing their importance in spring. “It’s time to trim perennials and grasses to promote new growth.” He describes these Spear & Jackson secateurs as “simple yet durable tools that excel in functionality.”
Photo: Jonathan Buckley
Niwaki Hori Hori Knife
£32.95 at Sarah Raven
When asked about their preferred tools, professional gardeners consistently mention the Hori Hori. It’s noted as a favorite among experts for summer preparations. “It’s multifunctional, ideal for digging, cutting, weeding, and planting,” says Lulu Roper-Caldbeck.
Spring Sweater
Borden Enid Embroidered Jumper
£112 at John Lewis
“I adore chic sweaters,” states Jess Kurtner Morley in April’s Style Essentials. “Finding a stylish jumper can be a challenge,” she remarks, highlighting its versatility for cool evenings. “I pair this with a black camisole for a layered look.”
Five-Star Rated Coconut Milk
Biona Organic Coconut Milk
£11.95 for 6 x 400ml on Amazon
“I prefer high-quality coconut milk; there’s a noticeable difference between products loaded with sweeteners and stabilizers,” remarks Ravinder Bhogal in a tasting of coconut milk.
Barrel Leg Trousers
Jersey Trousers
£34.90 at Uniqlo
Initially highlighted by Jess Cartner Morley as a cost-effective way to achieve a high-end look, these Uniqlo Barrel Leg trousers have earned popularity as one of our top choices in the Women’s Spring Wardrobe Essentials guide. Their wearability is certainly a selling point.
“Compact, efficient, and easily fits in your pocket,” praises reviewer Pete Wise about this 102g power bank. “No other power banks match its combination of charging prowess and portability,” he adds. With a 5,000mAh capacity, it provides ample energy for most smartphones.
The Best Sunscreen
Altruistic Sunscreen SPF50
£27.95 for 1 liter on Amazon
“Using sunscreen is essential; otherwise, anti-aging products are futile,” states Sari Hughes, an expert on age-defying skincare. This dermatologist-recommended formula might seem pricey, but it is an excellent investment in high-quality sun protection.
‘Glass Hair’ Conditioning Gloss
Glaze Clear Conditioning Gloss
£15.99 on Amazon
Hayley Spencer consulted experts on achievable, DIY-friendly beauty treatments. “Glass hair” can be easily attained at home. “The shine and freshness of salon-treated hair are undeniably appealing,” she mentions. This clear conditioning gloss, which can be applied to dry hair, takes only 10 minutes to work its magic.
Best Budget Walking Pad
Rattantree Shock Absorbent Treadmill
£142.49 at Debenhams
While there’s no substitute for walking outdoors, under-desk treadmills are on the rise for boosting daily steps while at work. The Rattantree model provides solid value at £142.49, as noted by Fitness Fanatic Leon Poultney. “It may lack the frills of pricier models, but it’s lightweight and covers the basics for anyone looking to increase their step count.”
Pollock-Style Bud Vase
Rainbow Vase
Starting at £12.50 on Etsy
A popular choice among readers, this vibrant vase adds a splash of color to any space, designed to hold a single flower stem. It’s a simple and affordable way to brighten your home. Fashion writer Ellie Violet Brumley notes that it pushes the boundaries of traditional splatter patterns.
Although diverse animal sequences have evolved the ability to use tools (primates, parrots, octopus, crabs, hornets, etc.), the factors leading to evolutionary use that lead to tools are less understood. Fish can provide insight into these factors by comparing differences between tool-used and non-used species. The use of anvils is an example of the use of tools by fish. The fish holds the densely packed prey in its mouth, attacking it on an anvil to open it. Through the community science programmes we call Using fish toolsMarine biologists have collected 16 new observations of five anvil use in a group of WRASSES (colorful coral reef fish) called the New World Harikoeles. These new observations provide the first evidence of the use of ANVIL Halichoeres brasiliensis, Halichoeres poeyi and Halicoeres radiatusand first video evidence of anvil use Halichoeres Garnotti and Halichoeres bivittatus.
Tool-used genus within the family Labridae. Image credit: Taliel Adam et al. , doi: 10.1007/s00338-025-02633-w.
The use of tools was once considered human characteristics and a fundamental role in human evolution.
Evidence suggests that the use of tools is widespread among animals.
Animals use tools when using external objects to accomplish a particular task.
The tool allows animals to accomplish tasks and do them easier. The use of tools appears to be extremely beneficial for animals, and raises the question of why all animals do not use the tools.
Previous research into the evolution of tool use has been limited by morphology and ecology, which contrasts with the difficulty of finding closely related populations or species of varying degrees of use of tools.
Although it has not been well studied, the use of tools in fish may be an exception. Fish are the most distinctive vertebrate groups with high ecological and morphological diversity, even among closely related species.
The use of anvils is an example of tool use in fish. With anvil, the fish (i) grabs hard shell prey, such as sea urchins and bivalves, in their mouths, (ii) swims on hard surfaces such as rocks and coral heads, and (iii) opens the fierce surface quickly and repeatedly.
The use of anvils is described in 26 fish species, and all WRASSEs belonging to the family Labridae.
“The use of tools is usually related to humans, but this behavior proves that fish are far more clever than trust,” says Dr. Juliet Taliel Adam, a researcher at Macquarie University.
Through the use of fish tools in the Citizen Science Initiative, Dr. Taliel Adam and her colleagues gathered 16 new observations in five species. Harikoeles WRASSES.
The findings present the first evidence of three anvil use and two other two video evidence, extending the known anvil use range into the Western Atlantic.
“With these newly discovered tool-used species, it becomes clear that many species of Huaras use tools they didn’t know before,” Dr. Taliel Adam said.
“This study adds to the study of fish intelligence,” added Callum Brown, a senior author at Macquarie University.
“They demonstrate the use of flexible and dexterous tools and are expanding their understanding of the evolution of tool use in the animal kingdom.”
Team’s result It will be displayed in the journal Coral reef.
____
J. Taliel Adam et al. Using tools by the New World Harikoeles WRASSES. Coral reefPublished online on March 26th, 2025. doi:10.1007/s00338-025-02633-w
The Mid-Lestic Age is considered a dynamic period in European and African history, but is generally considered a static period in East Asia. The discovery of a series of refined stone tools at Long Tang Sight in southwestern China, 50,000-60,000 years ago, challenges that recognition.
A product from the Quina system located at Longtan Site, China. Image credits: Luan et al. , doi: 10.1073/pnas.2418029122.
The Paleolithic period in the Central era occurred about 300,000-40,000 years ago and is considered an important time in human evolution.
This period is related to the origins and evolution of modern African people.
In Eurasia, it is associated with the development of several archaic human groups, such as the Neanderthals and Denisovan.
However, there is a widely believed belief that in most of the Paleostemic period, China’s development had slowed.
“Our discoveries challenge our current understanding of human history and technological development in East Asia,” said Professor Beau Lee of the University of Wollongong.
“This finding challenges a long-established general theory among archaeologists that China-China tools are relatively simple and unchanged.”
Professor Li and colleagues unearthed a rich collection of stone tools at the Longtan archaeological site in Yunnan, China.
This tool revealed a complete kina technology system that includes the cores used to generate large and thick flakes.
The Kina Industry is one of the most representative tool creation strategies developed in the Mid Paleolithic period around 300,000-40,000 years ago.
It is characterized by a steeply scaled retouch of thick flakes, which often produces robust scrapers with heavy edge modifications associated with neanderthals, representing strategies developed during marine isotope stage 4 as a response to open forest grassland environments and cool/dry climates.
The Kina Technical System was discovered in Western and Southern Europe during this period, but was not thought to have existed in East Asia.
The wear traces of Longtanquina scrape suggest that they are used in a variety of materials, including bones, horns, wood, meat, skin, and non-slow plants.
“Evidence has shown that the discoveries at Longtan significantly broadened the geographical distribution of human species, the tools used, and the adaptability to adapt to a variety of climates and environments,” Professor Li said.
“The discovery of Longtan also provides a perspective to understand how the human-leeze species evolved and evolved in East Asia prior to the massive arrival of early modern people 45,000 years ago.”
“Understanding rather old artifacts forces us to rethink models of human migration patterns and technology evolution in this part of the world.”
“This opens an exciting new avenue for research and, as we know, can rewrite East Asia’s prehistoric period.”
a paper The survey results were published this week Proceedings of the National Academy of Sciences.
____
Qi-Jun Ruan et al. 2025. Kinarithic technology demonstrates the diverse late Pleistocene human dynamics of East Asia. pnas 122 (14): E2418029122; doi: 10.1073/pnas.2418029122
A new online platform has been launched to help speakers practice in front of virtual audiences, easing the anxiety many face in public speaking situations.
Dr. Chris McDonald, founder of Cambridge University’s Immersive Technology Lab and creator of the platform, aims to eliminate long waits and high costs associated with seeking help for language anxiety.
“Most people experience language anxiety but don’t have access to treatment. This project aims to break down those barriers,” he explained.
Virtual reality public speaking The platform uses exposure therapy, combining breathing exercises and eye movements to reduce heart rate and fear response.
Users can practice public speaking in various virtual reality settings, from empty classrooms to large stadiums with thousands of people. McDonald refers to the latter as “overexposure therapy.”
McDonald mentioned that the platform, compatible with Android and iOS, offers scenarios like study materials, feedback mechanisms, and job interviews accessible via laptop, VR headset, or smartphone with a cheap mount.
In a recent study published in the Frontier Journal of Virtual Reality, 29 Chinese adolescents showed significant improvement in public speaking confidence and enjoyment after using the platform.
Further research is planned, but McDonald revealed that tens of thousands have already used the platform during development. He emphasized the importance of creating an effective and accessible tool for users.
Psychologist Dr. Matteo Cella from King’s College London’s Virtual Reality Lab acknowledged the platform’s potential benefits but stressed the need for robust trials to evaluate its efficacy.
Dr. Kim Smallman of Cardiff University highlighted the importance of assessing the impact and effectiveness of new technologies like VR in addressing mental health challenges.
Washington – Early Man utilized animal bones to create cutting tools 1.5 million years ago on a regular basis.
A recent discovery of 27 sculptures and sharp bones from elephants and hippos in Tanzania’s All-Bai Valley site has extended the timeline for the use of ancient bones by around a million years. Researchers already knew that early individuals crafted simple tools from stones as long ago as 3.3 million years.
New discoveries of ancient humans published in Nature on Wednesday have shown that they had a more sophisticated toolkit, incorporating various materials, according to William Harcourt Smith, a paleontologist at the American Museum of Natural History. He was not involved in the study.
A well-preserved bone tool measuring up to 16 inches (40 cm) may have been created by fracturing the bulky ends of leg bones and chipping off the flakes from the remaining bone shaft using stones. Research co-author Ignacio de la Torre, a researcher at the Spanish National Research Council, explained that this technique was used to produce one sharp edge and one tip.
The bone tools were likely used as hand axes, handheld blades not attached to a handle, for the purpose of butchering animal carcasses.
These blades were ideal for removing flesh from elephant and hippo carcasses but were not utilized as spears or projectiles. “I don’t believe they were hunting these animals. They were likely scavenging,” he stated.
Numerous artifacts exhibit signs of being struck in order to remove more than dozens of flakes, indicating a sustained level of craftsmanship.
The consistent choice of bones – specifically large, heavy leg bones from a particular animal – and a pattern of uniform modifications suggest that early humans deliberately selected and crafted these bones, as noted by paleobiologist Milia Pacheco from the Federal University of San Carlos in Brazil, who was not part of the study.
The bones show minimal signs of erosion, trampling, or gnawing by other animals, ruling out the possibility of natural factors shaping the tools.
These bone tools date back over a million years, predating the emergence of our species, Homo Sapiens, by approximately 300,000 years.
According to Brianna Poviner, a paleontologist with the Smithsonian Human Origins program, when the tools were created, three types of human ancestors coexisted in the same East African region.
This tool could have been created and used by Homo Erectus, Homo Habilis, or Paranthropus Boisei. “It could have been any of these three, but it’s nearly impossible to determine which one,” Poviner mentioned.
Envision the future of HR. Picture receiving a notification on your phone informing you that due to recent organizational changes, new personnel need to be recruited. The message includes a list of six well-qualified candidates who align with the organizational culture and are available to start within a month. Your task is simply to choose the best candidate to interview.
Much of this future scenario is already a reality. Platforms like Employment Heroes offer advanced AI tools to assist small businesses in managing HR and recruitment. These tools can even predict future needs and suggest potential candidates proactively.
Utilizing AI tools, Employment Heroes analyzes clients’ businesses, including organizational structure, turnover rates, and hiring timelines, making it easier for small businesses to operate. This global employment management platform uses AI to provide insights and recommendations, such as identifying when a position needs to be filled.
This workforce planning capability has traditionally been inaccessible to small and medium-sized businesses, placing them at a disadvantage in recruitment and staffing. Now, they have access to expertise and support equivalent to that of large corporations.
The right candidates can already line up as soon as a vacancy occurs. Photo: Maria Corniva/Getty Images
This type of predictive HR is particularly beneficial for high-growth SMEs. By providing insights into future staffing needs, it enables strategic planning that ensures the right talent is in place at the right time.
Looking ahead, AI-enabled HR platforms will automate various employment management processes, from employment terms to bonus structures, streamlining operations for small businesses and ensuring fairness and transparency.
Baillie, the Head of People at Mobile Marketing Agency ConsultmyApp, highlights the impact of the Employment Hero platform on recruitment processes, emphasizing its role in enhancing inclusivity and ensuring competitive pay packages.
SmartMatch enables SMEs to deliver competitive packages by analyzing industry trends. Photo: FG Trade/Getty Images
Employment Heroes’ SmartMatch feature offers real-time data insights on industry trends, helping small businesses set competitive pay ranges and attract top talent.
By leveraging data-driven benchmarks, small businesses can align their compensation packages with market standards and ensure they remain competitive in attracting and retaining the best employees.
Let’s reimagine the possibilities. Discover how Employment Hero can revolutionize your work processes.
Microsoft’s artificial intelligence models can replicate realistic video game footage the company says will help designers create games, but experts are not convinced that the tool will help most game developers. yeah.
Neural networks that can generate consistent, accurate footage from video games are nothing new. A recent Google-created AI produced a fully playable version of a classic computer game destiny No access to the underlying game engine. original destiny, However, it was released in 1993. More modern games are more complex with sophisticated physics and computationally intensive graphics, and have proven difficult for AIS to faithfully replicate.
now, Katja Hofmann Microsoft Research and her colleagues developed an AI model called Muse. This allows you to recreate the complete sequence of multiplayer online battle games Bleeding edge. These sequences follow the physics that underlie the game, and appear to keep players and in-game objects consistent over time. This means that the model has a deeper understanding of the game, says Hofmann.
The Muse is trained with seven years of human gameplay data, including both controllers and video footage. Bleeding EdgeNinja Studios is a Microsoft-owned developer. It works similarly to large language models such as ChatGpt. If given input, it imposes predicting the next gameplay in the form of video game frames and their associated controller actions. “To this day, for me, it’s a very moving thing to me, purely from training models to predict what will come next. I learn a sophisticated and deep understanding of this complex 3D environment,” Hoffman said. I say it.
To understand how people use AI tools like Muse, the team researched game developers and learned which features would be useful. As a result, researchers added the ability to repeatedly adjust changes made on the spot, such as changes to player characters or new objects entering the scene. This could help you come up with new ideas and try out what-if scenarios for developers, says Hofmann.
However, the muse is still limited to generating sequences within the original boundaries Bleeding Edge Games – Can’t come up with new concepts or designs. And I say it’s unclear whether this is a model-specific limitation or something that can be overcome with more training data from other games. Mike Cook King’s College London. “This is a long way from the idea that AI systems can design their own games.”
The ability to generate consistent gameplay sequences is impressive, but developers may prefer greater control, says Cook. “If you create a tool that is actually testing the game code itself, you don’t have to worry about persistence or consistency because you’re running the actual game. So these are introduced by generative AI itself. It’s solving the problem.”
It is promised that the model is designed with developers in mind, he says Georgios Yannakakis The Digital Games Institute at the University of Malta may not be feasible for most developers who don’t have that much training data. “Does that come down to the question of it being worth it?” says Yannakakis. “Microsoft has been collecting data for seven years and training these models to demonstrate what they can actually do. But real game studios can afford it. [to do] this? “
Even Microsoft itself is vague about whether AI-designed games could be on the horizon. When asked if there was a possibility that developers in the Xbox gaming division would use the tool, the company declined to comment.
Hofmann and her team hope that future versions of Muse can generalize beyond training data, but they can create new scenarios and levels for the games they are trained to work in a variety of games. I hope that I can do it. Challenge is because modern games are very complicated.
“One way games distinguish themselves is by changing the system and introducing new concept-level ideas. So machine learning systems go outside of their training data and go beyond what they see. It’s extremely difficult to innovate and invent,” he says.
Over the past two weeks, EPA staff have had to compete with dramatic reforms at the agency.
Approximately 1,100 “probation” employees were said to be possible to be fired, and 168 staff working on environmental justice issues have been taken on leave.
Lee Zeldin, new manager at the EPA, said agency priorities include improving AI and automotive jobs.
Lee Zeldin led the Environmental Protection Agency in a short and a half weeks. HR movements shaking wildly, rattling some staff, like many others.
On the day of Zeldin's confirmation last week, the EPA notified about 1,100 “probation” employees that they could close at any time for less than one year.
Then on Thursday, the agency placed 168 staff members on administrative leave. The affected people were working on environmental justice issues across the EPA's 10 regional offices and headquarters.
This week, the agency removed an online mapping tool called EJScreen. It was being used by federal, state and local governments to make decisions that support environmental justice. The term explains the idea that people should have equitable access to a clean, healthy environment, and that some underserved communities face historically unbalanced environmental harms. It refers to. For example, state highway agencies You can use EJScreen Check demographic information for which the road construction project was planned.
Zeldin took on his post a day after a federal worker received a “road fork” email offering resignation shopping. Their deadline to accept the offer was Thursday night, but a federal judge put the initiative on hold that day, following legal challenges from the union. The program will be blocked until at least Monday.
“We accused Congress of being as efficient as possible with the taxes sent to us,” Zeldin said, adding that Americans “are feeling a lot of financial pain.” Ta.
His initial actions and the shock they inflict on staff suggest that Zeldin and the Trump administration are not wasting time dramatically reworking the EPA and redefine its purpose. .
Environmental Protection Agency spokesman Molly Vaserio said the EPA is focused on adhering to President Donald Trump's executive orders, including an order entitled “Extreme and Waste Government DEI Program.” Ta.
“The EPA is enthusiastically implementing President Trump's executive order and subsequent related implementation memos. President Trump has been elected to delegate from the Americans to do this,” Vaselio said. .
Several EPA staff members said fear and hype quickly permeated the agency.
“The past two weeks have been pretty scary,” said Marie Owens Powell, chairman of the U.S. Government Employees Federation Council 238, the union representing around 8,500 EPA staff. “Every day, it was something. It was exhausting.”
Powell worked As an EPA Storage Tank InspectorAdded that there were other recent surprises, such as when the pronouns of staff preference were removed from email signatures without notice.
Another EPA worker asked that his name not be made public due to fear of retaliation, but explained that the feeling was “limbo” or “purgatory.”
“We are afraid of doing work that can be seen as completely opposed to the executive order or against Trump's agenda. We want to speak up and push back. But the fear is obvious,” the staff said. “We're all waiting to see who's next.”
Vaseliou met with staff at the Career EPA to visit several disaster sites, including East Palestine, Ohio, and visited several disaster sites, including trains carrying chemicals in February 2023, and to visit several disaster sites, and toxins. He said he had spent his first few weeks of emitting smoke. He also went to Los Angeles. There, a wildfire that broke out last month raided thousands of homes and headed to West North Carolina where Hurricane Helene killed dozens of people.
In Zeldin's news release on Tuesday We laid out five priorities For the EPA under his leadership, including an effort to “pursuing energy independence,” we will develop “the cleanest energy on the planet” to ensure clean air and water. However, some of his agenda diverges from the core mission of the EPA, at least as it operates under past administrations. These include advances in artificial intelligence, reforming and reviving permits for auto work.
Jeremy Simmons, senior adviser to the Environmental Protection Network, a former EPA staff group, said he is worried about the direction the agency is on the lead, based on Zeldin's statement.
“If you're worried about toxic contamination in your community, it's difficult to see yourself on that agenda,” said Symons, who worked at the EPA from 1994 to 2001, in the service of the political agenda. ”
Congressional Democrats appear to be preparing for the fight for the future of the EPA. D-Mass. Sen. Ed Markey of the group attempted to enter agency headquarters on Thursday, calling for a meeting with representatives from Elon Musk's Government Efficiency Bureau.
“We just went in and asked for a meeting with a representative from Doge. We were denied and we were turned away,” Markey said at a press conference outside the building.
Vaseliou said he had not taken the appropriate steps necessary to allow Markey to enter headquarters and described the event as a “promotion stunt.”
A Markey spokesperson said Thursday that the senator has not received confirmation as to whether Doge's representatives are at the EPA. However, multiple sources say the names of workers that NBC News identified as members of Doge member Cole Killian were listed in the EPA directory.
Emails to Killian's EPA email requesting an interview were not immediately returned. Vaseliou did not answer questions about Killian or whether he was connected to Doge.
When asked about Marquee's concerns on Thursday, White House deputy press secretary Harrison Fields said Democrats were “gaslighting” about Doge's mission.
“To reduce waste, fraud, abuse and become a better steward of the hard-earned dollars for American taxpayers may be a crime for Democrats, but it's not a crime in court,” Fields said. I said that.
The United Kingdom has become the first country to implement laws regarding the use of AI tools, as highlighted by a remarkable enforcement organization overseeing the use of this technology.
It is now illegal to possess, create, or distribute AI tools specifically designed to generate sexual abuse materials involving children, addressing a significant legal loophole that has been a major concern for law enforcement and online safety advocates. Violators can face up to five years in prison.
There is also a ban on providing manuals that instruct potential criminals on how to produce abusive images using AI tools. The distribution of such material can result in a prison sentence of up to three years for offenders.
Additionally, a new law is being introduced to prevent the sharing of abusive images and advice among criminals or on illicit websites. Border units will be granted expanded powers to compel suspected individuals to unlock and submit digital devices for inspection, particularly in cases involving sexual risks.
The use of AI tools in creating images of child sexual abuse has increased significantly, with a reported four-fold increase over the previous year. According to the Internet Watch Foundation (IWF), there were 245 instances of AI-generated child sexual abuse images in 2024, compared to just 51 the year before.
These AI tools are being utilized in various ways by perpetrators seeking to exploit children, such as modifying a real child’s image to appear nude or superimposing a child’s face onto existing abusive images. Victim voices are also incorporated into these manipulated images.
The newly generated images are often used to threaten children and coerce them into more abusive situations, including live-streamed abuse. These AI tools also serve to conceal perpetrators’ identities, groom victims, and facilitate further abuse.
Senior police officials have noted that individuals viewing such AI-generated images are more likely to engage in direct abuse of children, raising fears that the normalization of child sexual abuse may be accelerated by the use of these images.
A new law, part of upcoming crime and policing legislation, is being proposed to address these concerns.
Technology Secretary Peter Kyle emphasized that the country cannot afford to lag behind in addressing the potential misuse of AI technology.
He stated in an Observer article that while the UK aims to be a global leader in AI, the safety of children must take precedence.
Concerns have been raised about the impact of AI-generated content, with calls for stronger regulations to prevent the creation and distribution of harmful images.
Experts are urging for enhanced measures to tackle the misuse of AI technology, while acknowledging its potential benefits. Deleclehill, the CEO of IWF, highlighted the need for balancing innovation with safeguarding against abuse.
Rani Govender, a policy manager at NSPCC’s Child Safety Online, emphasized the importance of preventing the creation of harmful AI-generated images to protect children from exploitation.
In order to achieve this goal, stringent regulations and thorough risk assessments by tech companies are essential to ensure children’s safety and prevent the proliferation of abusive content.
During the Stone Age, the operation of stone tools provides evidence that it was not exclusive to humans, despite the popular image of early humans sitting around campfires or hunting. Even wild chimpanzees ( Pantosloid ) have been observed using stones to open nuts. They place a nut on a flat rock (called an anvil) and strike it with another stone (a hammer), or sometimes with a thick branch.
Research from Africa’s Tai National Park shows that chimpanzees have been using these tools for over 4,000 years. This suggests that the use of stone tools may be a trait inherited from common ancestors, although it’s also possible that different species learned this independently.
Other primates, like capuchin monkeys in Brazil and long-tailed macaques in Thailand, have also been observed using hammer and anvil technology. The capuchin monkeys take up to eight years to master the skill, using stones weighing up to 1 kg (2.2 pounds) to crack nuts, with evidence of this behavior dating back at least 3000 years.
While humans are known for using more advanced stone tools, animals also have their own versions – Credit: Mikroman6
Stone tools are not exclusive to primates either. Otters use stones to crack open shellfish and extract sea snails, while crows in New Caledonia drop hard nuts from heights onto anvils to open and access the contents.
Although stone tool use was once thought to be unique to Homo sapiens, archaeologists have found evidence of stone tool use by earlier human species like Homo habilis. Stone tool manufacturing remains a key aspect of human evolution, connecting ancient actions with the use of these tools. The discovery of other primates using stone tools challenges some of the oldest archaeological sites.
In 2022, archaeologists in Argentina proposed the idea that stone tools found in Brazil 50,000 years ago might have been created by capuchin monkeys, not humans. The quartz tools are similar to those created by present-day capuchin monkeys.
The use of stone tools by monkeys raises questions regarding the origins of tool use between humans and monkeys – CREDIT: chain45154
If these ancient tools were indeed made by monkeys, it would extend the timeline of their stone tool use by thousands of years and raise questions about when Homo sapiens arrived in South America.
Archeologists have also noted similarities between monkey-made stone fragments and human cutting tools, suggesting that early humans may have been influenced by accidental creations to develop their own tools. The discovery raises questions about the origins of the oldest stone tool artifacts.
This article answers the question, “Are we the only species to have passed the Stone Age?” posed by Juanita Andrade via email.
To submit your own question, please email questions@sciencefocus.com or reach out via the Facebook, Twitter, or Instagram page (don’t forget to include your name and location).
For more fascinating science, check out our ultimate fun fact page.
A new artificial intelligence tool known as Parlex can notify ministers about the potential unpopularity of a policy within their own party through a search called “parliamentary mood check.”
Parlex is just one of the AI tools being developed for ministers and civil servants to anticipate issues with backbenchers and pinpoint supportive legislators.
By inputting a policy outline like a 20 mph speed limit, the tool can predict how legislators will respond based on their past contributions in Congress. A demonstration video on the government website demonstrates historical opposition from Conservative MPs and support from Labour MPs for traffic calming measures.
Described as a “vibe check,” the tool helps policy teams understand the political landscape and develop response strategies before formally proposing a policy in Congress.
According to a report by The Times, key MPs like Iain Duncan Smith and former MP Tobias Ellwood oppose the 20mph limit, while Labour MP Kerry McCarthy supports traffic calming measures.
The tool is expected to be more beneficial for civil servants than ministers who should already possess a good understanding of congressional views.
Chancellor Keir Starmer recently announced an AI plan involving significant investments in Britain’s computing capacity to integrate the technology into the nation’s infrastructure.
The government’s initiatives include releasing public data to foster AI businesses, including anonymized NHS data for research and innovation purposes with strong privacy safeguards in place.
Ministers believe AI could stimulate Britain’s economic growth and generate an estimated economic boost of up to £470bn over the next decade.
Parlex is just one of many AI tools being developed within the government, with other tools like Redbox aimed at automating document analysis for civil servants.
The tool will soon be available to all civil servants in the Cabinet Office and DSIT, aiming to streamline manual processes and increase efficiency.
Another program called Consult generates revenue by automating consultation processes, allowing civil servants to better analyze and act on public opinion.
The Department for Work and Pensions has also utilized AI, including ‘whitemail’ to analyze letters received daily and communicate information more effectively.
However, challenges have arisen, such as inaccuracies in identifying housing benefit fraud suspects due to government algorithms underperforming.
Researchers at the University of Cambridge have found that artificial intelligence (AI) tools have the ability to influence online viewers into making decisions, such as what they purchase and who they vote for. The researchers from Cambridge’s Leverhulme Center for the Future of Intelligence (LCFI) are exploring the concept of the “intention economy,” where AI assistants can understand, predict, and manipulate human intentions, selling this information to companies for profit.
According to the research, the intention economy is seen as a successor to the attention economy, where social media platforms attract users with advertising. The intention economy involves technology companies selling information about user motivations, from travel plans to political opinions, to the highest bidder.
Dr. Johnny Penn, a technology historian at LCFI, warns that unless regulated, the intention economy will turn human motivation into a new form of currency, leading to a “gold rush” for those who sell human intentions. The researchers emphasize the need to evaluate the impact of such markets on free and fair elections, freedom of the press, and fair market competition.
The study highlights the use of large-scale language models (LLMs) in AI tools like ChatGPT chatbots, which can predict and guide users based on behavioral and psychological data. Advertisers in the attention economy can buy access to user attention through real-time bidding on ad exchanges or future advertising space on billboards.
In the intention economy, LLMs work with brokered bidding to leverage user data for maximum efficiency in achieving objectives, such as selling movie tickets. Advertisers can create customized online ads using generative AI tools, with AI models driving conversations across various platforms.
The research suggests a future scenario where companies like meta may auction off users’ intentions for activities like booking restaurants and flights to advertisers. AI models will adapt their output based on user-generated data, providing highly personalized formats. Tech executives have discussed the potential of AI models to predict user intent and behavior, highlighting the importance of understanding user needs and desires.
The probability that one word follows another can be used to create watermarks for AI-generated text.
Vikram Arun/Shutterstock
Google uses artificial intelligence watermarks to automatically identify text generated by its Gemini chatbot, making it easier to distinguish between AI-generated content and human-written posts. This watermarking system could help prevent AI chatbots from being exploited for misinformation and disinformation, as well as fraud in schools and business environments.
Now, the technology company says it is making available an open-source version of its technology so that other generative AI developers can similarly watermark output from their large-scale language models. I am. Pushmeet Kohli Google DeepMind is the company's AI research team, combining the former Google Brain and DeepMind labs. “SynthID is not a silver bullet for identifying AI-generated content, but it is an important building block for developing more reliable AI identification tools,” he says.
Independent researchers expressed similar optimism. “There is no known way to reliably watermark, but I really think this could help detect some things like AI-generated misinformation and academic fraud,” he said. I say. scott aaronson at the University of Texas at Austin, where he previously worked on AI safety at OpenAI. “We hope that other leading language modeling companies, such as OpenAI and Anthropic, will follow DeepMind’s lead in this regard.”
In May of this year, Google DeepMind announced Google announced that it has implemented the SynthID method for watermarking AI-generated text and video from Google's Gemini and Veo AI services, respectively. The company recently published a paper in the journal nature SynthID generally performs better than similar AI watermarking techniques for text. The comparison involved evaluating how easily the responses from different watermarked AI models were detectable.
In Google DeepMind's AI watermarking approach, as a model generates a sequence of text, a “tournament sampling” algorithm subtly moves it toward selecting “tokens” of specific words that are detectable by associated software. Create a statistical signature. This process randomly pairs candidate word tokens in tournament-style brackets. The winner of each pair is determined by which one gets the highest score according to the watermark function. Winners advance through successive tournament rounds until there is one round remaining. The “layered approach” “further complicates the potential for reverse engineering and attempts to remove watermarks,” it said. Yellow Furong at the University of Maryland.
It said a “determined adversary” with vast computational power could remove such AI watermarks. Hanlin Zhang at Harvard University. But he said SynthID's approach makes sense given the need for scalable watermarking in AI services.
Google DeepMind researchers tested two versions of SynthID that represent a trade-off between making watermark signatures easier to detect in exchange for distorting the text typically produced by AI models. They showed that the undistorted version of the AI watermark continued to work without noticeable impact on the quality of the 20 million text responses Gemini generated during live experiments.
However, the researchers also acknowledged that this watermarking works best on long chatbot responses that can be answered in a variety of ways, such as composing an essay or an email, as well as on math or coding questions. The response to this has not yet been tested.
Google DeepMind's team and others have stated the need for additional safeguards against misuse of AI chatbots, and Huang similarly recommended stronger regulation. “Requiring watermarks by law addresses both practicality and user adoption challenges and makes large language models more secure to use,” she says.
Archaeologist removing excess mud from a Bronze Age plow
wessex archeology
A Bronze Age wooden spade has been unearthed by British archaeologists. It is extremely rare that wooden artifacts from such an ancient period have been preserved.
This spade offers a glimpse into life at a time when people grew crops and increasingly lived in settled communities.
“It’s very specific,” he says ed treasure At Wessex Archaeology, Salisbury, UK. “It’s a very direct connection to the past.”
The spade was discovered in a bog near Poole Harbor on England’s south coast, where Wessex Archeology has been excavating for several years. of The Moors of the Arne Coastal Change Project is working to restore the region’s coastal wetlands, and archaeologists are carrying out excavations to ensure that no useful artifacts are accidentally lost.
The researchers were digging a ring gully, a circular trench that is thought to have originally surrounded the shelter. In one of the ring gullies they found the handle of a spade. “It was an almost unbelievable moment,” says Treasure, who was not personally present. “I could tell right away that it was a carved piece of wood.” The spade was carved from a single piece of oak.
Moist conditions did not expose the shovel to oxygen, slowing down rot.
Using debris found with the spade, the team radiocarbon-dates the spade to 3,400 to 3,500 years ago. “A small part of the spade broke off during the burial, but we used it as a date,” Treasure said. Nearby pottery also showed a similar age. This dates the origins of the spade back to the Middle Bronze Age.
“We’re in the midst of some pretty big changes in prehistoric Britain,” Treasure says. People stopped living as nomads and began spending more time in settled communities, growing a variety of grains and other foods.
However, there are no signs of permanent settlement at this location throughout the year. Of course, it was and still is a wetland. “We strongly believe this is a seasonal use of this landscape,” Treasure says. People may have brought in animals for summer grazing, cut peat for fuel, or collected reeds for thatching.
Future research will try to understand how plows were made and what they were used for. “It may have been used to cut peat on site,” Treasure said. “It may also have been used to dig the ring ditch where it was discovered.”
It is rare that a spade from this era has been preserved. One of the only other examples is the Brynlow shovel. Discovered in Cheshire in 1875rediscovered in the 1950s It was discovered by fantasy author Alan Garner in a school assembly hall and was eventually radiocarbon dated to almost 4,000 years ago.
Humpback whale (Megaptera novaegangliae) Whales build so-called “bubble nets” as tools to hunt, but researchers from the University of Hawaii and the Alaska Whale Foundation have found that these marine animals not only build bubble nets, but also manipulate these unique tools in a variety of ways to maximize food intake in their Alaskan feeding grounds. This novel study sheds light on a behavior critical to whale survival and makes a compelling case for including humpback whales among the rare animals that manufacture and use unique tools.
Although some animal species use tools to forage, only a few manufacture or modify tools. Humpback whales are one of these rare species, as they manufacture bubble net tools while foraging. Using animal tags and unmanned aerial system technology, Sabo others. Investigating the bubble nets made by solitary humpback whales (Megaptera novaegangliae) while feeding on krill in the Pacific Ocean, whale nets in southeast Alaska. They demonstrate that the nets are composed of internal tangential loops, suggesting that the whales actively control the number of loops, the size and depth of the net, and the horizontal spacing of adjacent bubbles. They argue that the whales adjust structural elements of the net to increase the amount of prey they catch in a single lunge by an average of seven-fold. Image credit: Szabo others., doi: 10.1098/rsos.240328.
“Many animals use tools to find food, but few actually make or modify their own tools,” says Lars Bader, a professor at the University of Hawaii.
“We discovered that solitary humpback whales in southeast Alaska build complex bubble webs to capture krill.”
“These whales expertly blow bubbles in patterns that form nets with internal loops, actively controlling details such as the number of loops, the size and depth of the net, and the spacing of the bubbles.”
“This method allows them to catch up to seven times as many prey in one feeding dive without expending any extra energy.”
“This impressive behaviour puts humpback whales in a rare group of animals that make and use their own tools to hunt.”
The marine mammals known as cetaceans include whales, dolphins and porpoises and are notoriously difficult to study.
Advances in research tools are making it easier to track and understand whale behaviour, and in this case, Professor Bader and his colleagues used specialised tags and drones to study the whale's movements from above and below the water.
“We attached non-invasive suction cup tags to whales in southeast Alaska and flew drones over humpback whales as they hunted alone in bubble nets to collect data on their underwater movements,” said Dr William Goff, also from the University of Hawaii.
“The tools are great, but it takes practice to hone them.”
“Whales are a difficult group to study and successful tracking with tags and drones requires skill and precision.”
“This little-studied foraging behavior is unique to humpback whales.”
“It's really amazing to see these animals in their natural habitat, doing things that only a handful of people get to see.”
“And then it's rewarding to be able to go back to the lab and look at the data and find out what they're doing underwater after they're out of sight.”
a paper The study was published in the journal Royal Society Open Science.
_____
A. Sabo others2024. Solitary humpback whales manufacture bubble nets as a tool to enhance prey intake. R. Soc. Open Science 11(8):240328;doi: 10.1098/rsos.240328
This article is based on a press release provided by the University of Hawaii.
A researcher plucks the feathers of a bird as part of an experiment to investigate Neanderthal cooking techniques.
Mariana Navaiz
To learn more about Neanderthal culinary talents, archaeologists cooked five wild birds using only fire, their hands, and stone tools. The experiment shows that our ancient relatives needed significant manual skill to use a flint blade to butcher an animal without injuring themselves.
Neanderthals lived in Europe and Asia until about 40,000 years ago. Hearths have been found at many of their sites, and there is evidence that they hunted large animals such as elephants and cave lions.
Mariana Navaiz Researchers at the Catalan Institute of Human Palaeoecology and Social Evolution in Tarragona, Spain, say that by recreating ancient activities such as cooking and butchering using tools available at the time, scientists can gain insight into how prehistoric humans lived.
She and her colleagues wanted to better understand archaeological bird remains associated with Neanderthals that date back about 90,000 years ago and were found in deposits in Portugal.
The team selected five birds that had died at a wildlife rehabilitation center in Portugal and were similar in size and species to those found at the archaeological site.Crow), Turtledove(Columba Palumbus) and two European collared doves (Streptopelia decaoctoThe tools used in the experiment were pieces of flint prepared by students of stone tool technology.
All five birds were plucked by hand. The crow and pigeon were butchered raw, while the remaining three were roasted over charcoal. The cooked birds were easily butchered without stone tools, but the raw birds required considerable effort using a flint blade.
“Paleolithic knives were certainly very sharp and required careful handling,” Navais says. “The precision and effort required to use these tools without injuring oneself highlights the practical challenges Neanderthals would have faced in their everyday food processing activities.”
Once the dissection was complete, the researchers prepared the bones and analyzed them for distinctive marks caused by stone tools and fire, as well as identifying wear marks from flint tools.
The burn marks and tool marks were then compared to Neanderthal food remains found at the archaeological sites of Fighiera Brava and Oliveira in Portugal, where bird bones with burn marks and cut marks matched the team’s reconstruction, Navaís said.
“Our experimental studies demonstrate that flaked raw birds display characteristic cut marks, especially around tendons and joints, while roasted birds display burn marks and increased brittleness leading to fractures,” she says. “These findings help distinguish between human-induced modifications and those caused by natural processes or other animals, such as trampling or the activity of rodents, raptors and carnivores.”
Neanderthals were skilled enough to capture and prepare small, fast-moving animals like birds, Navais said. “This study highlights the cognitive capabilities of Neanderthals and demonstrates their ability to capture and prepare small, fast-moving prey like birds, challenging previous ideas that they were incapable of such complex tasks.”
Sam Lin Researchers at the University of Wollongong in Australia say experimental archaeology is like reverse engineering, comparing what happens in modern samples with archaeological material to try to interpret what happened in the past.
In this case, one of the main findings is that cooked birds don’t require tools to prepare them for eating, which means some bones may not necessarily bear tool marks. “They learned that you just need to tear apart a cooked wild bird, just like we do when we eat barbecued chicken,” Lin said.
A convicted sex offender who created over 1,000 indecent images of children has been forbidden from using any “AI creation tools” for the next five years, marking a significant case in this realm.
Anthony Dover, 48, was instructed by a British court in February not to use artificial intelligence-generated tools without prior police authorization, as part of a sexual harm prevention order issued in February.
The prohibition extends to tools like text-image generators that produce realistic-looking photos from written commands, as well as the manipulation of websites used to generate explicit “deepfake” content.
Mr. Dover, who received a community order and a £200 fine, was specifically directed not to utilize the Stable Diffusion software known to be exploited by pedophiles to create surreal child sexual abuse material.
This case is part of a series of prosecutions where AI-generated images have come to the forefront, prompting warnings from charities regarding the proliferation of such images of sexual abuse.
Last week, the government announced the creation of a new crime that makes it illegal to produce sexually explicit deepfakes of individuals over 18 without their consent, with severe penalties for offenders.
Using synthetic child sexual abuse material, whether real or AI-generated, has been illegal under laws since the 1990s, leading to recent prosecutions involving lifelike images produced using tools like Photoshop.
These tools are increasingly being used to combat the dangers posed by sophisticated synthetic content, as evidenced by recent court cases involving the distribution of such images.
The Internet Watch Foundation (IWF) emphasized the urgent need to address the production of AI-generated child sexual abuse images, warning about the rise of such content and its chilling realism.
Law enforcement agencies and charities are working to tackle this growing trend of AI-generated images, with concerns rising about the production of deepfake content and the impact on victims.
Efforts are underway to address the growing concern over AI-generated images and deepfake content, with calls for technology companies to prevent the creation and distribution of such harmful material.
The decision to restrict adult sex offenders from using AI tools may pave the way for increased surveillance of those convicted of indecent image offenses, highlighting the need for proactive measures to safeguard against future violations.
While restrictions on internet use for sex offenders have existed, limitations on AI tools have not been common, underscoring the gravity of this case and its implications for future legal actions.
The company behind Stable Diffusion, Stability AI, has taken steps to prevent abuse of their software, emphasizing the importance of responsible technology use and compliance with legal guidelines.
This website uses cookies so that we can provide you with the best user experience possible. Cookie information is stored in your browser and performs functions such as recognising you when you return to our website and helping our team to understand which sections of the website you find most interesting and useful.
Strictly Necessary Cookies
Strictly Necessary Cookie should be enabled at all times so that we can save your preferences for cookie settings.
