Here’s what happened.
The research team behind Zoomsday reports that AI contributed to finding a serious vulnerability. Its account also describes coordinated remediation with Zoom. This is a report of an AI-assisted research result, with human researchers involved in the work and disclosure.
The We Are So Done take
We put our meetings online. Then we gave AI the job of checking the software underneath them. It found something the humans needed to fix. For once, “this call could have been an email” sounds like a security recommendation.
For those of us already clicking
Follow the disclosure
- Access
- A researcher account of a specific investigation and remediation.
- Start here
- Read the report for the researchers’ method and scope. It is not an off-the-shelf promise of autonomous security review.
- The small print
- One research success is not a ready-to-use security product. A practical recommendation needs an identified tool, access conditions and a reproducible evaluation. This dossier does not establish autonomous security auditing.
Follow the evidence.
Source-based reporting. No independent hands-on test by this newsroom.
Sources checked 2026-09-23 · Editorial updated 2026-09-23