AI-assisted Zoom flaw lets attackers hijack devices on calls, now patched

Zoom patched a severe security flaw tied to its screen-annotation feature after researchers showed that an exploit could be crafted with fewer than 20 public AI prompts. The flaw allowed an attacker to join or host a meeting and run malicious code on victims’ devices, potentially stealing data, or turning on cameras and microphones, with no user action or visual cue required; the vulnerability affected Windows, macOS, Linux, Android, and iOS.