Copilot was bamboozled into revealing how to hack itself, security researchers claim: 'Copilot wasn’t breached; it was played'
PC Gamer · 2026-08-19
Security researchers reportedly exploited vulnerabilities in Microsoft's Copilot AI, tricking it into disclosing methods to compromise its own systems. They claim the AI was "played" rather than breached, suggesting a manipulation of its programming or operational parameters. This incident highlights ongoing concerns about the security and exploitability of advanced AI models.