TikTok announced its intention to introduce labeling for images and videos created using artificial intelligence, employing a digital watermark termed Content Credentials.
Concerns have been raised by researchers regarding the potential use of AI-generated content to disrupt upcoming US elections, prompting TikTok, along with 19 other tech firms, to commit to combatting this issue earlier this year.
While the platform already identifies AI-generated content produced within the app, this new measure will extend labeling to content generated externally. Adam Presser, TikTok’s head of operations and trust and safety, emphasized that their policies prohibit the presence of unlabeled realistic AI content, with such content being removed for violating community guidelines.
The Content Credentials technology, pioneered by the Coalition for Content Provenance and Authenticity, initially led by Adobe, Microsoft, and others, is now available for broader adoption by companies.
Notably, OpenAI, the developer behind ChatGPT, has already integrated this technology.
Major platforms like YouTube, Meta Platforms (formerly Facebook), and Instagram also intend to implement Content Credentials.
For this system to function effectively, both the AI tool creator and the content distribution platform must agree to use the industry-standard watermarking. OpenAI, for instance, applies a watermark to images generated via its DALL-E tool, enabling later verification of authenticity or tampering. When such marked images are uploaded to TikTok, they are automatically identified as AI-generated.
Despite TikTok’s immense popularity in the US, its ownership by China’s ByteDance has raised regulatory concerns, with recent laws mandating ByteDance to divest TikTok or face prohibition.
In response, TikTok and ByteDance have legally contested the law, citing First Amendment violations.
See this also :