Announced on "Good Morning America" Thursday, the social media giant revealed its new policy to automatically label Artificial Intelligence-generated content when uploaded from select platforms.
"Our users and creators are enthusiastic about AI and its creative potential," said Adam Presser, TikTok's Head of Operations & Trust and Safety, in an exclusive interview with ABC News. "At the same time, we want to ensure people can distinguish between fact and fiction." TikTok asserts it is the first video-sharing platform to adopt Content Credentials technology, an open standard enabling tracing the origin of various media types.
"This is akin to a nutrition label for content, detailing the image's context, location, creator, and editing history," explained Adobe's Chief Trust Officer Dana Rao in an earlier interview with ABC News. Adobe is a founding member of the Coalition for Content Provenance and Authenticity, promoting the adoption of this digital standard. Content Credentials (digital nutrition labels) are increasingly used to certify digital content.
OpenAI recently announced plans to incorporate this technology into all images created and edited by DALL.E 3, its latest image model. OpenAI also intends to integrate Content Credentials into its video-generational Model, Sora, upon its broader release. Other products with generative AI capabilities, like Adobe Firefly, Photoshop Express, and Microsoft Copilot, are already utilizing the technology to embed metadata into visual content.
"The puzzle pieces are coming together," commented Sam Gregory, executive director of Witness and deepfake expert. "It's crucial for companies to make it easy to identify content created with their tools by providing tool-specific classifiers." TikTok announced that the new label rollout begins immediately and will apply to all users globally in the coming weeks. Additionally, Content Credentials will soon be attached to content, remaining with the content upon download to allow other platforms to access the metadata.
With the U.S. presidential election looming and elections worldwide, the demand to detect AI-generated content online is growing.
"Neither is a cure-all, now or in full deployment; they are a form of harm reduction that makes it easier to discern AI use," explained Gregory. "Malicious creators will still find a way around them, but most of us will be glad to use them if they don't compromise our privacy or creativity."
Rao envisions an online future where content with these digital nutrition labels may be more trustworthy or valuable than those without.
"The true value lies in authenticity," emphasized Rao. "Creators want authentic communication with their audience, and these credentials allow them to establish that level of authenticity."