AgentsThe story, in brief
A Zoom Screen-Sharing Bug Let Anyone Take Over Other Devices on a Call
Fewer than 20 prompts. That's all it took for an AI tool to find a critical Zoom vulnerability that let attackers hijack devices on calls.
The KeyNews take
Why it matters
AI-powered security research is finding real-world exploits at scale, but also raising the bar for vulnerability discovery — and the risks when those tools are weaponized. This is both a validation of AI's capability and a cautionary tale about autonomous attack surface.
The key facts
5 to knowPublic AI tool found Zoom screen-sharing RCE flaw in <20 prompts
Vulnerability allowed device hijacking of call participants
Flaw now patched by Zoom
Researchers demonstrated AI-assisted exploit discovery
Illustrates dual-use risk: same AI techniques for defense and attack
Go to the source
Wired AIwired.com
Publisher excerpt: Researchers say it took fewer than 20 prompts for a public AI tool to find a flaw (now fixed) allowing anyone on a Zoom call to hijack another participants’ device.