Zoom promises to only train AI models with user data after opt-in

Zoom promises that it will not train AI models on audio, video and chats without explicit permission from users. The company came under fire last week when it amended its terms and conditions to make this possible.

Zoom responds in a blog post to the commotion that arose earlier. The company adjusted its general terms and conditions, which have been active since July 27. In it, Zoom writes that it can use customer data to train machine learning or artificial intelligence models. “Zoom may access, use, collect, create, modify, distribute, process, share, maintain and store service-generated data for any purpose, including for purposes such as machine learning and artificial intelligence, including training and updating algorithms and models,” it said. can be read in the new privacy policy. Earlier this year, Zoom released tools such as Zoom IQa smart assistant that the company creates in collaboration with OpenAI.

The changes were already implemented in March this year, but only became noticeable recently. Zoom is now responding to the commotion that arose as a result of those changes. The company says it is now responding to be more transparent about what data it collects. Zoom reports that it the terms and conditions, specifically section 10.4, has been re-amended. The company added that it “does not use audio, video or chats to train artificial intelligence models without consent.”

In practice, this means that users must give permission themselves via an opt-in. For business users, the administrators must do this. In group conversations, all participants will be notified of this.