Researchers Used AI to Build a Zoom Attack in Under 24 Hours

Security researchers used AI prompts to build an attack that exploited a Zoom screen-sharing flaw in under 24 hours. The flaw, now fixed under bulletin ZSB-26015, let attackers take full control of a device during a call with no action from the victim. It affected unpatched Zoom Workspace apps on Windows, macOS, iOS, Android, and Linux. In this author's view, this shrinking gap between finding and weaponizing flaws is an alarming trend for defenders.