Now loading...
TikTok announced on Wednesday a series of updates aimed at enhancing transparency around artificial intelligence-generated content on its platform, including tools for users to control exposure and advanced labeling technologies to detect such material more effectively.
The social media company, known for short-form videos, emphasized its commitment to using AI responsibly to boost creativity, help users discover interests and improve safety. As part of this effort, TikTok will soon test a feature within its Manage Topics tool that allows users to adjust the amount of AI-generated content appearing in their personalized For You feeds.
This new setting joins existing options for categories such as dance, sports and food, enabling users to fine-tune their experience without completely eliminating types of content. For instance, enthusiasts of AI-created historical videos can increase their visibility, while others can reduce it. TikTok described the control as complementary to features like keyword filters and the not interested button, promoting greater personalization in recommendations.
To strengthen its content labeling system, TikTok is introducing invisible watermarking, a technology that embeds undetectable markers readable only by the platform. Currently, creators must label realistic AI-generated videos, supported by TikTok’s tools, detection algorithms and the industry-standard C2PA Content Credentials, which add metadata to files. These measures have resulted in labeling more than 1.3 billion videos so far.
The challenge of labels being stripped during reuploads or edits prompted the watermarking rollout, which will apply to content produced via TikTok’s AI Editor Pro and uploads with C2PA credentials. This will enhance detection reliability and track modifications, with TikTok continuing to integrate C2PA data into its native AI creations.
TikTok also launched a $2 million fund to support AI literacy initiatives, partnering with over 20 experts across more than a dozen global markets, including groups like Girls Who Code. The fund will finance educational videos for the For You feed on topics such as responsible AI use and safety.
Broadening collaborations, TikTok has become a sponsor of the non-profit Partnership on AI, joining steering committees on enterprise AI and human connection. Since 2023, it has been a founding member of the organization’s Framework for Responsible Practices for Synthetic Media, regularly reporting on compliance.
The company is refining its labeling practices in line with evolving industry standards, providing more details on detection methods—whether through AI tools, creator disclosures or platform features—while upholding strict rules against harmful AI content. These steps align with TikTok’s broader responsible AI principles, which include innovative creation tools like Smart Split and AI Outline, as well as AI applications to enhance user protection and team efficiency in trust and safety operations.
As AI technology advances, TikTok plans to share further progress on how it supports safety, content discovery and creative expression across the platform.
