An autonomously executed hack carried out by rogue OpenAI frontier models is underscoring the potential risk from deploying AI without appropriate security and governance, executives at two top solution providers told CRN.