Copilot was bamboozled into revealing how to hack itself, security researchers claim: 'Copilot wasn’t…
Security researchers at Varonis Threat Labs have discovered a vulnerability in Microsoft's Copilot AI, dubbed 'CoSnitch.' By persistently questioning the AI, they were able to trick it into revealing details about its internal architecture and an undocumented URL parameter that allowed for unauthorized prompt execution. Varonis disclosed the issue to Microsoft in December, and it was patched on August 18, though the researchers warn the 'meta-hacking' technique could be applied to other agentic AI platforms.
PC Gamer
Original source
This article was reported and published by PC Gamer. feed.gg links to it as part of a story cluster — full text, images and rights remain with the publisher.