AI-generated YouTube videos for kids spread dangerous misinformation
A new investigation has found that many AI-made YouTube videos for kids are spreading false or risky information, and could even mess with how young viewers learn and think.
These videos often have strange stories or unsafe ideas, making experts worry about their impact on children.
Some videos show children doing dangerous things
Some of these videos show things like children riding without seat belts and walking in the middle of the road, calling US states by made-up names like "Ribio Island," or showing babies eating whole grapes and honey, stuff that can actually be dangerous for little ones.
AI content is now a big part of YouTube
AI-generated content now makes up a big chunk of YouTube's feed; one report estimated about 21% of YouTube's overall feed is shoddy AI content, and a separate investigation of Shorts recommended after watching children's channels found over 40% contained AI-generated visuals, and some channels post dozens every day.
Nearly half of recommended Shorts also use AI visuals with confusing messages about what's "educational."
Experts warn this could hurt kids' learning
Experts say these mixed-up messages can make it harder for kids to learn cause and effect.
As one professor put it, this is a "monster problem" that could slow down healthy brain development.
YouTube is trying to fix the problem
YouTube has started labeling realistic-looking AI videos, but most cartoons slip through without any warning.
That means lots of misleading animated clips still get shown to kids without anyone knowing they're made by AI.