Researchers Built Working Zoom Exploit in Under 24 Hours Using AI Prompts

Security researchers used AI prompts to build a working exploit for a Zoom screen-sharing vulnerability in under 24 hours. The flaw, patched under bulletin ZSB-26015, allowed full remote device takeover during calls with no victim interaction across all unpatched Zoom Workspace clients on Windows, macOS, iOS, Android, and Linux. In this author's view, the 24-hour exploit timeline signals that defenders' traditional discovery-to-weaponization lag is eroding.