Researchers from A Security discovered vulnerabilities in Zoom's screen-sharing annotation protocol that allow silent device takeover without any victim interaction. The flaws were found in under 20 AI prompts using publicly available models, affecting all supported operating systems, and Zoom has begun rolling out fixes.
Background
Zoom is one of the world's most widely used video conferencing platforms, with hundreds of millions of daily users relying on it for work and personal communication. The rise of AI-assisted vulnerability discovery is increasingly lowering the barrier for finding software flaws.
- Source
- Ars Technica
- Published
- Aug 12, 2026 at 09:37 PM
- Score
- 7.0 / 10