OpenAI is dedicated to providing copyright holders with “greater control” over character generation following the recent release of the Sora 2 app, which has overwhelmed platforms with videos featuring copyrighted characters.
Sora 2, an AI-driven video creation tool, was launched last week by invitation only. This application enables users to produce short videos from text prompts. A review by the Guardian of the AI-generated content revealed instances of copyrighted characters from shows like SpongeBob SquarePants, South Park, Pokémon, and Rick and Morty.
According to the Wall Street Journal, prior to releasing Sora 2, OpenAI informed talent agencies and studios that they would need to opt out if they wished to prevent the unlicensed use of their material by video generators.
OpenAI stated that those who own Guardian content can utilize a “copyright dispute form” to report copyright violations, though individual artists and studios cannot opt out of blanket agreements. Varun Shetty, OpenAI’s Head of Media Partnerships, remarked:
OpenAI Sora 2 Generated Video 1
On Saturday, OpenAI CEO Sam Altman stated in a blog post that the company has received “feedback” from users, rights holders, and various groups, leading to modifications.
He mentioned that rights holders will gain more “detailed control” as well as enhanced options regarding how their likenesses can be used within the application.
“We’ve heard from numerous rights holders who are thrilled about this new form of ‘interactive fan fiction’ and are confident that this level of engagement will be beneficial for them; however, we want to ensure that they can specify the manner in which the characters are utilized.”
Altman noted that OpenAI will “work with rights holders to determine the way forward,” adding that certain “generation edge cases” will undergo scrutiny within the platform’s guidelines.
He emphasized that the company needs to find a sustainable revenue model from video generation and that user engagement is exceeding initial expectations. This could lead to compensating rights holders for the authorized use of their characters.
“Creating an accurate model requires some trial and error, but we plan to start soon,” Altman said. “Our aim is for this new type of engagement to be even more valuable than revenue sharing, and we hope it’s worth it for everyone involved.”
He remarked on the rapid evolution of the project, reminiscent of the early days of ChatGPT, acknowledging both successful decisions and mistakes made along the way.
Source: www.theguardian.com
Related Posts
