AI chatbots have failed people in crisis. Can that be fixed?

Ars Technica作者： Cyrus Farivar 2026年8月7日正文已收录本站

Clinicians and researchers say AI companies need to open up their safety data.