Researchers disclosed Zoom vulnerabilities that could have allowed an attacker on a screen-sharing call to silently take control of another participant’s device, without requiring any action from the victim. The flaws matter because Zoom is widely trusted for work, personal calls and public events, potentially making the attack a route into corporate networks as well as individual devices.

The weaknesses were in Zoom’s real-time screen-sharing annotation protocol and affected supported Windows, macOS, Linux, iOS and Android devices. Digital defence firm A Security said it found and developed an exploit for the bugs in early June using publicly available AI models in fewer than 20 prompts; Zoom has since issued server-side and client-side fixes.