💬 Viewpoint: Evaluating mental health #AI safety requires measuring both acute and cumulative harms, not just comparing specialized vs general-purpose models, and demands transparent, outcome-based standards. ja.ma/3QzWX57

1 likes 0 replies