TikTok will begin automatically labeling AI-generated content created on other platforms, the company announced Thursday. With this change, when a creator posts content on TikTok that was created using a service like OpenAI’s DALL·E 3, it will automatically have an “AI generated” label added to it to inform viewers that it was created using AI .
The social video platform achieves this by implementing Content Credentials, a technology from the Coalition for Content Provenance and Authenticity (C2PA), co-founded by Microsoft and Adobe. Content credentials attach specific metadata to content, which TikTok can then use to instantly recognize and label AI-generated content.
As a result, TikTok will automatically begin labeling AI-generated content that is uploaded to the platform with content credentials attached. The change will roll out on Thursday and will apply to all users worldwide in the coming weeks.
While TikTok already labels content created with TikTok AI effects, it will now also label content created on other platforms that have implemented content credentials, such as. B. DALL·E 3 from OpenAI and Bing Image Creator from Microsoft. While Microsoft, Adobe and OpenAI already support content credentials, Google has pledged to also support content credentials.
Although TikTok already requires creators to disclose when they post content created or enhanced with AI, the company told TechCrunch that it sees the new change as an additional way to ensure that AI-generated content is labeled, and at the same time taking the labeling pressure from the creators.
In the coming months, TikTok will also begin attaching content credentials to AI-generated content created on the platform using TikTok AI effects. Content credentials metadata contains details about where and how the AI-generated content was created or edited and remains associated with the content as it is downloaded. Other platforms that adopt content credentials can automatically mark the content as AI generated.
While TikTok is committed to labeling AI content on its own service, the company also wants to help ensure that AI content created in TikTok is properly labeled even when published on another platform.
“AI-generated content is an incredible creative outlet, but transparency for viewers is crucial,” said Adam Presser, Head of Operations and Trust & Safety at TikTok, in a press release. “By working with colleagues to label content across platforms, we are making it easier for creators to engage with AI-generated content responsibly while continuing to deter harmful or misleading AIGC, which is banned on TikTok.”
TikTok touts that it is the first video sharing platform to implement content credentials. It’s worth noting that Meta announced back in February that it plans to build on the C2PA solution to add provenance to content.
As part of Thursday’s announcement, TikTok said it is committed to combating the use of deceptive AI in elections and that its policies strictly prohibit harmfully misleading AI-generated content – regardless of whether it is labeled or not.