[Researchers Discover Zoom Vulnerabilities with Fewer Than 20 AI Prompts, Build Attack Chain Within 24 Hours] According to Decrypt, Israeli cybersecurity company ASecurity released a report stating that a researcher used publicly available AI models to discover multiple vulnerabilities in Zoom's annotation tool with fewer than 20 prompts and built an exploitable attack chain within 24 hours. The vulnerabilities are identified as CVE-2026-53413, CVE-2026-53414, and CVE-2026-53415. Exploiting these vulnerabilities, attackers can remotely execute code during meetings without any action from the victim, gaining control of their devices, stealing data, or activating their microphone or camera.