Zoom rewrites its policies to make clear that your videos aren’t used to train AI tools

Zoom has updated its terms of service and reworded a blog post explaining recent terms of service changes referencing its generative AI tools. The company now explicitly states that “communications-like” customer data isn’t being used to train artificial intelligence models for Zoom or third parties. What is covered by communications-like? Basically, the content of your videoconferencing on Zoom.

Here’s the key passage from the newly-revised terms:

Zoom does not use any of your audio, video, chat, screen sharing, attachments or other communications-like Customer Content (such as poll results, whiteboard and reactions) to train Zoom or third-party artificial intelligence models.

Zoom has come under scrutiny over terms of service language that people interpreted as giving the company broad control and copyrights of customer data, including potentially anything they showed or discussed during a call, with the intention of using that content to train the AI models that power features like its meeting summaries.

Section 10 of the terms of service, where the previous language had been, has also been rewritten to more clearly differentiate between “customer content” and “service generated data.”

According to Zoom, the refreshed policy is just reiterating its position more definitively after the company already revised the terms earlier this week to try and assuage customers. Before Friday’s revision, the most recent version said, “Zoom will not use audio, video or chat Customer Content to train our artificial intelligence models without your consent,” without as clearly laying out what it would use or how consent is given.

On Monday, Zoom’s chief product officer Smita Hashim updated her blog post about these changes to state that “our customers continue to own and control their content.” The older version of Hashim’s blog post has more examples of how Zoom uses customer data to “provide value-added services,” that doesn’t include training its own models, like having a license to deliver the recording of a meeting and use an automated scanner that detects fraud or spam messages.

Still, this wasn’t the first time people raised concerns over seemingly-broad terms of service changes, and I think there’s more Zoom could have done here initially to clarify this for users. With growing fears about how companies are using data to train AI models, it definitely won’t be the last.