# AI benchmarking

AI Often Mistakes Polite Words for Safe Suicide Intervention

A new study shows AI chatbots frequently misjudge suicide intervention, favoring responses that merely sound...