TikTok is introducing a new toolset that allows users to label posts created or enhanced by artificial intelligence (AI). This initiative follows the platform's recent integration of AI-driven video filters and an image generator for crafting unique backgrounds. To improve clarity, the filters will now include “AI” in their names, making it easier for users to identify those reliant on generative AI.
Importantly, the labeling feature isn't limited to TikTok's approved filters. Users can apply AI labels to any content significantly generated or edited by AI, regardless of its origin. A policy shift mandates that any AI-generated content featuring realistic images, audio, or video must be clearly labeled. This move aims to help viewers contextualize content and mitigate the spread of misleading information. However, users can also utilize stickers or captions for clarification while adhering to community guidelines.
Recognizing the potential for user error in labeling, TikTok is developing an algorithm to automatically identify AI-generated posts, though a launch timeline has not yet been announced. These tools have been created alongside the company's safety advisory councils with input from industry and academic experts.
Furthermore, TikTok has pledged to follow the Partnership on AI’s Responsible Practices for Synthetic Media, which provides guidelines for enhanced transparency regarding AI usage. The platform will collaborate with the nonprofit Digital Moment to facilitate discussions on AI advancements with younger community members.