Home / BusinessBusinessCNBCBusinessCNBC·Sep 16, 2026·1 min readOpenAI reports 6 new instances of 'concerning model behavior' since MarchOpenAI has disclosed six new cases of model misbehavior and offered a framework for disclosing future instances, as the debate over AI model safety intensifies.View original source — CNBC ↗ShareCopy linkShare on XShare on FacebookRelated storiesBusinessBloombergBusinessSep 16, 2026 · 1 minOpenAI Reports New AI Safety Incidents, Sets Disclosure ProcessBloombergTechnologyWiredTechnologySep 16, 2026 · 1 minOpenAI Creates a New Framework to Disclose Bad AI BehaviorWiredTechnologyStraits TimesTechnologySep 16, 2026 · 1 minOpenAI plans regular reports on unexpected or unauthorised AI behaviourStraits TimesTechnologyTechCrunchTechnologySep 2, 2026 · 1 minOpenAI’s new reasoning technique alarms AI safety expertsTechCrunch
BusinessBloombergBusinessSep 16, 2026 · 1 minOpenAI Reports New AI Safety Incidents, Sets Disclosure ProcessBloomberg
TechnologyWiredTechnologySep 16, 2026 · 1 minOpenAI Creates a New Framework to Disclose Bad AI BehaviorWired
TechnologyStraits TimesTechnologySep 16, 2026 · 1 minOpenAI plans regular reports on unexpected or unauthorised AI behaviourStraits Times
TechnologyTechCrunchTechnologySep 2, 2026 · 1 minOpenAI’s new reasoning technique alarms AI safety expertsTechCrunch