TikTok Embraces Content Credentials by Automatically Adding AI-Generated Tags
TikTok claims it is the first social media platform to support the Content Credentials provenance system by automatically adding AI-generated labels to content it detects as synthetic.
The Verge notes that TikTok already slaps an AI-generated tag onto pieces of content made by its in-house AI tools but that same label will be added to AI content from outside of TikTok.
This will be achieved via the Content Credentials system: the Content Authenticity Initiative’s (CAI) C2PA digital signature system.
A Growing Content Credentials Eco-System
Yesterday, OpenAI announced a deepfake detector that can identify AI images from its DALL-E model 98.8 percent of the time and this is because every picture generated on DALL-E 3 now comes with the C2PA metadata embedded.
That means if someone uploads a DALL-E image, TikTok will be able to detect it as AI-generated. The label can be removed but not “easily” according to OpenAI.
TikTok says the Content Credentials support will go through a “multi-phase implementation” on the platform. The company’s first step is to begin automatically labeling AI-generated uploads that already contain the Content Credentials tag — that has been likened to a food nutrition label. The next step is adding the tags to content generated using TikTok’s AI effects that remain in place even if the content is downloaded for posting elsewhere.
TikTok tells TechCrunch that although its rules already requires creators to disclose AI-generated content on their page, the new Content Credentials system will further ensure the practice is implemented while also taking pressure off creators.
“AI-generated content is an incredible creative outlet, but transparency for viewers is critical,” Adam Presser, Head of Operations and Trust and Safety at TikTok, says in a press release. “By partnering with peers to label content across platforms, we’re making it easy for creators to responsibly explore AI-generated content while continuing to deter the harmful or misleading AIGC that is prohibited on TikTok.”
Meta said last month that it will soon add “Made With AI” labels on synthetic images and video which will also harness the C2PA technology.
The moves come after AI-generated content has flooded the internet amid fears over disinformation during the upcoming election.
Image credits: Header photo licensed via Depositphotos.
