A Zoom vulnerability discovered with the help of an AI tool allowed attackers to remotely execute code on devices used by meeting participants, including iPhones and Macs. According to WIRED, researchers found flaws in the video conferencing platform that could be exploited to take over targets’ devices during calls involving screen sharing. The attack required no victim interaction and left no trace. It affected all operating systems supported by Zoom‑Windows, macOS, Linux, iOS, and Android. Cybersecurity firm A Security disclosed the vulnerability to Zoom, and it has since been patched. The concerning aspect was how easily AI tools identified the flaw. Discovered in early June using publicly available AI models, the bug was uncovered with fewer than 20 prompts, enabling researchers to develop a working attack quickly. A Security co-founder Omer Gull told WIRED that such tasks previously required teams of five people and months of work. He emphasized that Zoom is a critical target because users inherently trust the platform. WIRED’s Lily Newman reported on the story, noting that Gull briefed her via a video call on Microsoft Teams.

The incident highlights growing concerns about AI-assisted vulnerability discovery and its implications for software security. As AI becomes more accessible, malicious actors could potentially exploit similar weaknesses at scale. Zoom has addressed the issue with a patch, but the episode underscores the need for proactive security measures in widely used applications.

Organizations are urged to update their Zoom clients immediately to avoid potential exploitation. Security experts recommend enabling automatic updates and reviewing permissions for screen sharing and remote access features.

A Security’s discovery demonstrates both the power and risk of AI in cybersecurity. While helpful for identifying flaws before bad actors do, it also lowers the barrier for finding and weaponizing vulnerabilities.