TikTok is taking new steps to combat misinformation by automatically labeling AI-generated content.
The video-sharing app takes the step as the platform faces the possibility of being sold or banned in the United States.
Previously, TikTok only labeled AI content created with its own tools. We are currently working with the Coalition for Content Provenance and Authenticity (C2PA) to use a technology called Content Credentials. This technology allows platforms like TikTok to automatically identify and label AIGC.
The new feature focuses on identifying images and videos that were created with AI tools outside of TikTok. TikTok will soon launch a system to detect AI-generated audio content.
“Although AI presents great creative opportunities, it can be confusing or misleading to viewers if they don’t know that the content was generated by AI.Labeling It helps clarify that context. That's why we've been labeling AIGC created with TikTok AI effects and requiring creators to label realistic AIGC for over a year. “We also built the first tool to make it easy, and more than 37 million creators have used it since last fall,” TikTok said.
TikTok is also participating in the Adobe-led Content Authenticity Initiative (CAI) to promote more transparent content labeling practices in the industry. TikTok was the first video sharing platform to use content credentials.
TikTok is also focusing on educating its users. We work with media literacy organizations like MediaWise and WITNESS to create resources and launch educational campaigns. These campaigns aim to increase user awareness of AIGC.
TikTok will release 12 videos throughout the year in partnership with MediaWise, a program of the Poynter Institute. These videos teach universal media literacy skills and explain how his TikTok tools, such as AIGC labels, can help you understand content.
TikTok is also launching a campaign to raise awareness about AI labeling and the potentially misleading AIGC through a series of videos developed with WITNESS.
TikTok says that while most people want to enjoy AI-generated content responsibly, some may use AIGC to intentionally deceive others. TikTok is wary of these risks, and its policies strongly prohibit misleading AI-generated content, regardless of whether it's labeled or not. They also signed an industry agreement this year to combat the use of deceptive AI in elections.
TikTok further states, “While most people want to enjoy AI-generated content responsibly, there may be malicious actors who use AIGC to intentionally deceive others. We are alert to these risks, which is why our policies strictly prohibit harmful and misleading AI-generated content, whether labeled or unlabeled, and this year we Signed an industry agreement to combat the use of AI.”
