TikTok is getting tired of AI slop and is testing out new AI content guidelines 

The upcoming tweak would change TikTok’s detection algorithms, with AI-generated content on topics that can influence public trust and well-being coming under a more watchful eye.

While the platform did define which ones they are (as mentioned above), it did not explain how aggressive or effective this would necessarily be at cleaning up the platform, nor did it provide exact deadlines beyond “in the coming weeks”.

What other AI changes are there?

TikTok recently joined the C2PA Steering Committee. While it’s mainly a self-regulating club that promotes responsible, verifiable use of AI, its members are mostly organisations that would likely profit from AI use. Among its Steering Committee members are Meta, Google, Microsoft, and OpenAI — these get to decide on policy.

Brands like Qualcomm, Arm, Canon, Fujifilm, Leica, Logitech, Huawei, and many more are counted among its general members, with the understanding that they would implement these policies in their products and services, adhering to guidelines. An example is how Qualcomm’s Snapdragon processors for phones are C2PA-compliant, which we’ve reported on, as far back as 2020.

While is largely a shift in its governance and influence on  AI use, it’s worth noting that TikTok already watermarks AI content in an invisible manner (three billion of them, so far). Joining C2PA as a policy-shaping member would also mean this watermarking won’t go away anytime soon.

In-app hub for learning how to spot AI content

Separately, TikTok is now working on educating its users about AI-generated content. While no specific dates or mechanics were given, it said that the app will soon have a resource hub for users to learn “practical skills” in spotting AI content “when they search for AI-related terms”.

Source: TikTok (newsroom)