Why is the proliferation of AI-generated content harming the internet unchecked? – Arwa Mahdawi

HWhat do you think, humans? My name is Arwa and I am a genuine member of this species homo sapiens. We are talking about 100% real people; meat space This is it. I am by no means an AI-powered bot. I know, I know. That's exactly what the bot says, isn't it? I think you'll just have to trust me on this matter.

By the way, the reason I have such a hard time pointing this out is because content created by real humans is becoming kind of a novelty these days. The internet is rapidly being overtaken by advances in AI. (It's not clear who coined the term, but “slop” is a sophisticated iteration of Internet spam: low-quality text, video, and images generated by AI.) recent analysis It is estimated that more than half of all English long-form posts on LinkedIn are generated by AI. Meanwhile, many news sites are secretly experimenting with AI-generated content, in some cases signed. Author generated by AI.

Slop is everywhere, but Facebook is actively sloshing strange AI-generated images, including bizarre depictions. Jesus was made of shrimp. Much of the AI-generated content is created by fraudsters looking to drive user engagement, rather than remove them from their platforms. fraudulent purpose – Facebook accepted it. A study conducted last year by researchers at Stanford and Georgetown found that Facebook's recommendation algorithm is accelerating. These AI-generated posts.

Meta also creates its own slops. In 2023, the company began introducing AI-powered profiles like Liv, a “proud black queer mom of two and truth teller.” These didn't get much attention until Meta executive Connor Hayes talked about them. financial times The company announced in December that it plans to fill its platform with AI characters. I don't know why he thought bragging that soon we'll have a platform full of AI characters talking to each other would work, but it didn't. Meta quickly deleted the AI ​​profile after it went viral.

For now, people like Liv may be gone from Meta, but our online future looks increasingly sloppy. The gradual “ensitization” of the Internet, as Cory Doctorow memorably called it, is accelerating. Let's pray that Shrimp Jesus will perform a miracle soon. we need that.

Source: www.theguardian.com

The Proliferation of Fake AI Images Persists – 8 Notable Examples | Science & Technology Updates

Fact-checkers highlighted some notorious examples of AI-generated images that went viral this year, such as Prince William and Prince Harry embracing at the royal coronation.

Midjourney OpenAIWith DALL-E 3, you can now create realistic images faster and easier than ever using only text prompts.

While being a proponent of the technology known as generative; artificial intelligence,please tell me can empower artistsleading to concerns. Possibility of spreading false information.

Charity Full Fact has selected eight examples from 2023 that have been shared thousands of times.

They have since been marked as AI-generated or removed by social media platforms.

Prince William and Prince Harry reunite

Slideshow of 8 images showing prince of wales and the Duke of Sussex king’s coronation spread widely to Facebookover 78,000 likes!

In one of the photos, they appear to be hugging each other with teary eyes, but none of the photos are real.

According to a Full Fact investigation, these photos were originally published in a blog post in which the author explained how to use Midjourney’s image generator to “imagine a heartfelt reconciliation” between two people.

Julian Assange goes to prison

Photo of WikiLeaks founder The scenes at Belmarsh Prison were created using Midjourney.

The creator confirmed as much in an interview with Germany’s Bild newspaper, but not until the image was shared on Facebook and reposted 29,000 times. X.

Donald Trump’s portrait

Before the former US president posted a photo of his real face on Xmany fake versions were circulating.

Some of them have been viewed more than a million times, even though the jumble of letters behind him is a major feature. AI generators often have a hard time recreating text within images.

Mr Trump He had previously been the subject of an AI-generated image that appeared to show the moment of his arrest.

President Emmanuel Macron during the French riots

meanwhile riots in franceThe photo of has become a hot topic Emmanuel Macron He was sitting in the street as garbage burned behind him.

The image was widely shared, with one post garnering more than 55,000 views and comments suggesting the media was ignoring the story, according to Full Fact.

Pope Francis’ large audience

Photo of pope His speech to a large crowd in Lisbon was viewed tens of thousands of times on social media.

But a closer look revealed that it wasn’t real. One hand of the Pope had three fingers.

It comes months after an eerily convincing AI image of the Pope wearing a down jacket went viral.

Elon Musk’s “Robot Wife”

SpaceX Billionaire He makes no secret of his desire to create humanoid robots.but not “The Robot Wife.”

A post featuring an image of him kissing one such model was created by a digital artist and shared on Facebook and X.

Titanic submarine wreckage

While searching for titan submersibleMidjourney was used to create an image purporting to show debris.

It showed a game controller floating in the water, with the caption: “Breaking news: Exploded Titanic submarine controller found floating near the surface.”

The submarine is controlled using a modified controller and has been reported to have been sighted over 300,000 times on the X.

Rishi Sunak’s Bad Pint

image:
Image posted by Karl Turner MP (L) and original photo posted to Number 10’s Flickr account (R)

critic of prime minister The shot of him pouring out bad beer was an example of how he was portrayed as an out-of-touch person.

This image is a compilation of photos from the beer festival in August. Pint’s appearance worsened and onlookers looked confused.

It received over 78,000 views on X. The fact that Labor MP Karl Turner shared it also didn’t help.

Full Fact said the government and regulator Ofcom must prioritize public media literacy ahead of the next election, helping them recognize fake images and question what they see online. Ta.

Chief Executive Chris Morris added: “Failure to take action risks reducing people’s trust in what they see online. This risks undermining democracy, particularly during elections. Yes,” he added.

Source: news.sky.com