Alignment Hub beta launches with proposed minimum template for AI incident reports

Elie Bakouch and Thom Wolf announced the beta Alignment Hub and proposed a short template for AI incident disclosures. Wolf puts its collection at more than 120 publicly reported incidents and estimates that detection dates and monitoring details are often missing from reports; his post gives no denominator for those estimates. The creators also want to connect related incidents. Wolf says human review of AI-extracted entries continues.