Back to feedArs Technica - All contentTechnologynewsAI chatbots have failed people in crisis. Can that be fixed?Friday, August 7, 2026 Cyrus Farivar View original––– Clinicians and researchers say AI companies need to open up their safety data.
