TikTok to automatically label AI-generated content to combat misinformation

TikTok announced on “GMA” Thursday morning that beginning immediately, the social media giant will automatically label Artificial Intelligence-generated content when it is uploaded from certain platforms.

“Our users and our creators are so excited about AI and what it can do for their creativity and their ability to connect with audiences.” Adam Presser, TikTok’s Head of Operations & Trust and Safety told ABC News in an exclusive interview. “And at the same time, we want to make sure that people have that ability to understand what fact is and what is fiction.”

The social media giant says they are becoming the first video-sharing platform to implement Content Credentials technology — an open technical standard providing publishers, creators, and consumers the ability to trace the origin of different types of media.

“This is like a nutrition label for content. It tells you what happened in the image, where it was taken, who made it, and the edits that were made along the way,” explained Adobe’s Chief Trust officer Dana Rao during an interview with ABC News earlier this year.

Adobe is one of the founding members of the Coalition for Content Provenance and Authenticity, a coalition of companies working together to push forward the adoption of this new digital standard.

And Content Credentials (digital nutrition labels) are becoming more widely used as a standard to certify digital content.

Earlier this year, OpenAI announced that they would be adding this technology to all images created and edited by DALL.E 3, their latest image model. OpenAI also said it plans to integrate Content Credentials for its video-generational Model, Sora, once it is launched more broadly.

Other products with generative AI capabilities like Adobe Firefly, Photoshop, Express and Microsoft Copilot are among others who are already using the technology to embed metadata into visual content created using their platforms.

“Parts of the picture are falling into place,” remarked Sam Gregory, executive director of Witness and expert on deepfakes. “It’s essential that specific companies make it as easy as possible to know when content was created with their tools by providing tool-specific classifiers.”

TikTok said the rollout of this new label starts today and will apply to all users globally in the coming weeks. Over the coming months, TikTok will also start attaching Content Credentials to content, which will remain on the content when downloaded allowing other platforms to read the metadata.

With the United States presidential election around the corner and elections stacked across the world, demands to detect AI-generated content online have been growing more intense.

“Neither is a silver bullet, not now or when they are fully utilized and provided – they are a form of harm reduction that makes it easier to discern when and how AI was used,” explained Gregory. “Malicious and deceptive creators and distributors will still find a way around them, but most of us will be happy to use them if they don’t compromise our privacy or our ability to create.”

Rao believes that we will move toward an online world where content that carries these digital nutrition labels might be more trustworthy or more valuable than those that do not.

“The real value is going to be in authenticity,” emphasized Rao. “The creators want to have an authentic communication with their viewers, and so this content credential allows creators to establish the level of authenticity that they want with their audience.”