TikTok will start tagging AI generated content!

TikTok will begin labeling content created using artificial intelligence and uploaded from external platforms to combat the spread of misinformation.

"AI brings incredible creative opportunities, but if viewers don't know that content is generated by AI, it can cause confusion or mislead," TikTok said in a statement on Thursday. "Labeling helps clarify this context—that's why we label AIGC (AI-generated content) created using TikTok AI effects, and for more than a year we've required creators to label authentic AIGC."

TikTok's policy change is part of a broader effort in the tech industry to provide more safeguards for the use of AI.

In February this year, Meta announced it was working with industry partners to develop technical standards to make it easier to identify images generated by AI tools, and eventually expand to the identification of video and audio. At that time, users on Facebook and Instagram will see labels on AI-generated images.

Google stated last year that AI labels would be launched on YouTube and other platforms.

As for TikTok, it is also working with the Coalition for Content Provenance and Authenticity and using its content credential technology.

TikTok says this technology can attach metadata to content, then use the metadata to instantly identify and label AI-generated content. It also stated that it has begun deploying this technology on images and videos, and will apply it to pure audio content in the future.

In the coming months, content on TikTok will be attached with content credentials, and the credentials will remain on the content after downloading. This will help identify AI-generated material made on TikTok and help people understand when, where, and how the content was created or edited. Other platforms adopting content credentials will also be able to automatically label such content.

"Using content credentials as a direct way to identify and communicate synthetic media to audiences is an important step toward AI transparency, and is even more meaningful than typical watermarking technology," said Claire Leibowicz, head of the AI and Media Integrity Program at the Partnership on AI, in a statement. "At the same time, we need to better understand how users react to these labels, and we hope TikTok will report feedback so we can better understand how the public responds to an increasingly AI-enhanced world."

TikTok says it is the first video sharing platform to put content credentials into practice, and will join Adobe-led content authenticity initiatives to promote the adoption of credentials in the industry.

TikTok's past policy was to encourage users to label content generated or heavily edited by AI. It also requires users to label all AI-generated content containing real images, audio, and video.

In response, Adam Presser, TikTok's Head of Operations and Trust & Safety, said in an interview with foreign media: "We want to make sure people can understand what is fact and what is fiction."