Cybersecurity researchers have identified a critical vulnerability in Zoom. The flaw can allow an attacker to take control of another user’s device during a live meeting. It is linked to Zoom’s annotation feature within the screen-sharing system.
The issue is a memory-corruption bug. It affects Zoom clients on Windows, macOS, iOS, Android, and Linux in versions before 7.1.5. Clients using end-to-end encryption in version 7.0.6 also remain vulnerable. Earlier reports noted that versions up to and including 7.0.5 were affected.
Once exploited, an attacker can join or host a meeting and target any participant. The takeover requires no action from the victim. It also produces no visible warning on the compromised device. After gaining control, the attacker can steal personal data, activate the microphone or camera, or install additional malicious software.
Researchers at A Security discovered the vulnerability. They stated that they developed a working exploit using fewer than 20 prompts with publicly available AI models. The entire process, including finding the flaw and creating the exploit, took less than 24 hours according to their account.
The team reported the issue to Zoom in June. Zoom has acknowledged the report and released fixes. The vulnerabilities are tracked as CVE-2026-53413, CVE-2026-53414, and CVE-2026-53415.
Zoom advises users to protect themselves by installing the latest available updates. Applying the current version removes exposure to the known flaw.