YouTube's Algorithm Is Feeding Kids Bizarre AI Slop

YouTube has a kids problem. A New York Times investigation analyzed over 1,000 YouTube Shorts videos recommended to children and found the algorithm aggressively surfaces bizarre, often nonsensical AI-generated content.

The findings are grim. Many of these AI-generated videos contain conflicting information and lack any disclosure that they were machine-made. Experts warn the low-quality content flooding kids' feeds isn't just annoying — it's potentially harmful.

The platform's recommendation engine appears to have no meaningful filter distinguishing quality children's content from AI-generated noise. Once a child engages with one such video, the algorithm happily serves up more.

YouTube has faced years of criticism over its kids content pipeline. But the explosion of cheap generative AI tools has made the problem dramatically worse. Anyone can now flood the platform with synthetic content at scale — and the algorithm will do the rest.