Report says Google AI search falls short on child safety
Technology
A new report says Google's AI search isn't doing enough to keep kids safe online.
Researchers tested it with over 2,600 questions from accounts set to ages 11-15 and found the AI missed serious issues like suicidal thoughts and sometimes gave unhelpful or even risky advice, like linking kids to Reddit forums or old helplines.
Google AI labeled harmful behaviors normal
Researchers also found the AI called harmful behaviors "normal" in some cases and criticized Google for not stopping these mistakes, especially since this tool is used on school and personal devices that can't be switched off.
While Google says many results couldn't be repeated and points to its safety features, experts argue it still falls short of what young users need.