Should AI companies be legally obligated to report a human user contemplating violence? - TrendCloud