Zoom deepfake risk detection: what it means for avatar video

Zoom's Workplace announcement, read 2026-10-08, describes deepfake risk detection that gives real-time alerts when synthetic audio or video is detected, with a follow-the-user model that applies in hosted and external meetings. The same announcement lists realistic and stylized avatars that mirror a user's expressions and lip and eye movements, and avatars in Zoom Clips that turn slide decks into presentation clips. Those items were listed with expected availability of March and April 2026; confirm rollout on your account.

If you play a rendered avatar clip into a meeting, the page does not say how detection treats declared synthetic media. Plan for an alert and label the clip.

What the announcement lists

The wording below is from the announcement and the What's New at Zoom page. Neither says how the detector works, what it flags, or whether a host can mark a file as an approved synthetic source.

Designing for a detector you cannot inspect

Tell the audience before the clip plays: say in the meeting that the next video is an AI-generated presenter. Put the same line in the title of the shared file. If the clip is for external guests, send the file ahead of time rather than playing synthetic video in a meeting with security alerts enabled.

Sume Avatar 1.0 clips are fully synthetic by design. The docs describe a script-driven avatar, with captions optional, so a caption saying the presenter is AI is something you can add with inline captions or write into the first spoken line.

Cost of an approved briefing clip

A 30-second briefing with the disclosure line in the first sentence is $5.52 on Standard, $7.35 on Plus and $16.50 on Max. Render once, share the link, and avoid depending on a meeting detector at all. The disclosure checklist lists wording that fits.