Control Your TikTok AI Feed
▼ Summary
– TikTok is introducing a new setting that lets users control how much AI-generated content appears in their “For You” feed.
– The AI content control is part of the “Manage Topics” tool, allowing users to adjust their feed preferences without removing content entirely.
– TikTok is testing “invisible watermarking” technology to better label AI-generated content and prevent removal of labels during reuploading.
– The platform will apply invisible watermarks to AI content created with TikTok tools and content uploaded with C2PA’s Content Credentials.
– TikTok announced a $2 million AI literacy fund for experts to create educational content about AI literacy and safety.
TikTok is introducing a new feature that gives users direct control over the amount of AI-generated content appearing in their personalized “For You” feed. This significant update, integrated into the existing “Manage Topics” tool, allows individuals to fine-tune their content discovery experience rather than completely blocking specific types of media. The platform is simultaneously rolling out more sophisticated labeling technologies to improve transparency around synthetic media.
The AI-generated content (AIGC) setting functions similarly to the controls for over ten other categories, such as Dance or Food & Drinks. According to the company, the purpose is to help people shape the variety of content they encounter. It is designed to adjust the frequency of AI content visibility, not to eliminate it from feeds entirely.
This strategic move occurs as other tech giants are leaning into AI-centric content streams. Recently, Meta launched “Vibes,” a dedicated feed for short AI-generated videos. Shortly after, OpenAI introduced “Sora,” a social platform built specifically for creating and sharing AI video content. On TikTok, the presence of highly realistic AI videos has grown, with many creators also using AI tools to produce visuals for posts on subjects like history or celebrity culture.
The new control empowers users to personally curate their feed. Those who prefer less synthetic content can reduce its presence, while enthusiasts can opt to see more. To access this feature, navigate to Settings, select “Content Preferences,” and then choose “Manage Topics.” Inside, a slider for AI-generated content lets you easily adjust your preference. This update will become available to all users over the coming weeks.
To bolster its labeling system, TikTok is testing an “invisible watermarking” technology. The platform already mandates labeling for realistic AI-generated content and utilizes an industry-standard technology called Content Credentials from the C2PA coalition. This system embeds metadata into files to signal their AI-generated origin to platforms. A known limitation is that these labels can be stripped away when content is downloaded, re-uploaded, or edited elsewhere.
The new invisible watermarks are designed as an additional safeguard. They apply a digital mark that only TikTok’s systems can detect, making it significantly more difficult for third parties to remove. Initially, these watermarks will be applied to AI content created with TikTok’s own tools, like AI Editor Pro, and to any content uploaded to TikTok that already carries C2PA Content Credentials. The company has stated it will continue to both read existing C2PA credentials and attach them to AI content produced on its platform.
In a related initiative, TikTok announced the creation of a $2 million AI literacy fund. This fund is intended to support experts and organizations, such as the nonprofit Girls Who Code, in developing educational content that improves public understanding of AI, focusing on literacy and safety practices.
(Source: TechCrunch)
