TikTok is rolling out a toolset that lets users label posts that have been created or enhanced by artificial intelligence. This move comes after the social media giant added a number of filters for video uploads that made heavy use of AI, and an image generator to help create unique backgrounds. The filters are being renamed to make it clearer which ones rely on generative AI to further assist with labeling. Moving forward, these filters will have “AI” in the name somewhere.
The new labels aren’t exclusive to TikTok-approved filters. You can slap the label on any content that’s been completely generated or significantly edited by AI, no matter where the content has been sourced from. Disney+ could have used these tools back when Secret Invasion premiered.
TikTok’s new labeling tools aren’t exactly optional. A recent company policy clearly states that any AI-generated content that contains “realistic images, audio or video” must be labeled as such to “help viewers contextualize” the situation and “prevent the potential spread of misleading content.” However, you can use other disclosures, like stickers or captions, to get the point across without violating any community guidelines.
When you place the onus on consumers to label their content, mistakes are gonna happen. To that end, TikTok’s developing its own algorithm that will label AI-generated posts automatically, though it hasn’t announced when these in-house tools will launch. All of the aforementioned tools were developed in tandem with the company’s safety advisory councils, with assistance from industry and scholastic experts.
Additionally, the platform has committed itself to abiding by the Partnership on AI’s Responsible Practices for Synthetic Media, a list of recommendations for increased transparency involving artificial intelligence. TikTok will also partner with the nonprofit Digital Moment to host roundtables to discuss the advancement of AI with “young community members.”