Global AI Safety Incident Reporting Platform AISafetyWatch Launches: Real-Time Tracking of Actual Harm Caused by AI Systems

Global AI Safety Incident Reporting Platform AISafetyWatch Launches: Real-Time Tracking of Actual Harm Caused by AI Systems

The AISafetyWatch global AI safety incident reporting platform, jointly operated by the World Economic Forum and MIT, officially launched on July 31. The platform collects, verifies, and analyzes actual harm incidents caused by AI systems worldwide — from autonomous driving accidents to AI misdiagnoses, from algorithmic discrimination to AI-generated disinformation.

On its first day, the platform cataloged 342 verified AI safety incidents from the past 12 months, with medical AI misdiagnoses accounting for the highest proportion (31%), followed by autonomous driving accidents (22%) and algorithmic discrimination (19%). Each incident is accompanied by detailed technical analysis, root cause assessment, and improvement recommendations.

AISafetyWatch adopts a "no-blame" mechanism similar to the aviation industry's safety reporting system — encouraging companies and individuals to voluntarily report AI safety incidents without fear of legal repercussions, thereby promoting industry-wide safety learning. Over 200 tech companies have already registered with the platform's incident reporting system.