AIWS sets four levels for AI danger alerts

Boston Global Forum and AIWS launched the AIWS Trust Observatory with four alert levels, from Watch to Critical Alert, for monitoring consequential AI power. Its founding framework says the independent civil-society body will grade evidence, log public safety commitments, map disclosure blind spots, and publish its first quarterly State of Frontier AI at the end of January 2027. Builders should treat it as an external accountability signal, not a deployment control plane: its named-action reporting complements pre-deployment evaluation governance when teams need an auditable record of who must act on observed model risk.