AI chatbots have failed people in crisis. Can that be fixed?Ars Technica · LC · trust 39/100Clinicians and researchers say AI companies need to open up their safety data.Read the original at Ars Technica →Open in TruthVane →
