It may not be just an AI company that allows AI to access social media products that only have AI-generated content. After Meta announced the atmosphere a few days ago, Openai reportedly will soon release Tiktok, an AI-generated video slop. According to Wired (via Gizmodo), the app will feature the next-generation SORA 2 video model, which will be announced soon by the ChatGPT maker.
Openai released its first public version of SORA in December last year, long before tools like Google's VEO 3 arrived. However, SORA did not see the popularity experienced by Openai's 4o image generation model several months later. Meanwhile, Google's VEO 3 is one of the best AI video generation products released this year.
Openai's release of a social app focused on AI's only content is a bit worrying given that many people aren't grateful for AI slops. But that's not the only bad news. Openai is reportedly offering rather disturbing features to AI social apps, prompting users to verify their identity through facial recognition systems. The Wall Street Journal report also shows that copyright holders need to opt out of copyrighted works featured in SORA 2 content.
How apps like Openai's Tiktok work
Wired explains that the SORA 2 app will ask consumers to use facial recognition systems to verify their identity. The confirmed portrait can be used in SORA 2 AI videos. This sounds like what you'd expect from an AI-centric social network focused on video content. However, some users may not be used to checking their identity, especially those who don't want to create their own AI content.
It has even more controversial features tied to identity verification. Other users on the platform can use your portrait in AI Slops. You can still receive notifications even if the content is not out of the draft folder. It is unclear whether users can opt out of letting others use their faces in AI videos.
Otherwise, the Openai Tiktok clone should work like a social media app featuring short form video content. The app supports comments, likes, likes and content remixes, as well as showcasing Sora 2 videos up to 10 seconds in length. Users are reportedly able to create Sora 2 videos via text, photos and video prompts.
How about copyright protection?
According to Wired, SORA 2 is trained to avoid copyright violations when creating content. It shows whether real life use only works, but this is a useful protection. Limitations in AI programs can be bypassed by clever prompts. That said, the WSJ report claims that you can use copyrighted content to create videos unless SORA 2 opts out of appearing in SORA 2's work.
Openai was able to launch the Sora 2 on Tuesday, September 30th, so Openai appears to be informed about the opt-out process last week. Instead, the talent agency should report any violations it finds. This can turn out to be boring and continuous work. However, Openai has an agreement with the studio to prevent certain copyrighted characters from appearing in AI content.
The report also points out that SORA 2 does not generate images of famous public figures without permission. The new tool also requires the same copyright protection as the ChatGPT image generation tool, which went viral earlier this year. As a reminder, the product allowed users to create AI photos using Studio Ghibli anime style. Users were also able to generate deepfakes that featured famous celebrities' faces relatively easily, at least in the early days, at least after the launch of the ChatGPT 4o image generation tool.
