TikTok Introduces User Controls for AI Content

TikTok is introducing user controls that allow users to manage the amount of AI-generated content in their feeds. According to The Guardian, this change is being tested and will soon enable users to open the manage topic setting on the app and select preferences for AI-generated content.

Currently, TikTok hosts over 1.3 billion videos labeled as AI-generated, which is part of a broader increase in AI content driven by new tools like OpenAI's Sora and Google's Veo 3. Jade Nester, TikTok's European director of public policy for safety and privacy, noted that while many users enjoy AI-generated content, others may prefer less of it.

The initiative comes after reports indicated that nearly one in ten of the fastest-growing YouTube channels features only AI-generated videos, often categorized as low-quality or nonsensical, a trend referred to as 'AI slop.' TikTok's new feature will allow users to customize their experience, similar to existing options for topics like current affairs and beauty.

Additionally, TikTok is implementing guidelines requiring creators to label realistic AI-made content and is banning harmful deepfakes. Any AI video that is not labeled can be removed under its community policies, and the app will now add an AI watermark to content generated by its tools or flagged by the C2PA initiative for identifying AI content.

Furthermore, TikTok is launching a $2 million fund aimed at promoting AI literacy through partnerships with organizations like Girls Who Code, focusing on responsible AI usage. Amid ongoing concerns regarding content moderation, TikTok is also facing criticism for job cuts in its trust and safety team, with plans to lay off 439 moderators in London.

Despite this, TikTok asserts that human moderation will still play a crucial role, emphasizing the importance of balancing technology and human oversight in content moderation. Brie Pegum, TikTok's global head of program management for trust and safety, stated that AI systems have led to a 76% decrease in exposure to shocking content for human moderators, supporting the argument that AI can be beneficial in filtering harmful material quickly.