UNIO ANALYTICS

Incidents in which AI systems took unauthorised or harmful action that no person directed, including capability precedents with no offender to charge.