Zoom Responded to Its AI Policy Controversy
San Jose, CA – Over the weekend, an online user by the name of “isodev” posted on the ycombinator forum that Zoom’s Terms of Service allows AI training on user content without an opt-out option.
The Chief Product Officer, Smita Hashim, attempted to allay concerns in a blog post on Monday asserting that “We do not use audio, video, or chat content for training our models without customer consent.”
Zoom also tweeted on Monday morning to stress its commitment to transparency and user control.
In March, Zoom made changes to its Terms of Service, explicitly stating its right to employ user data for training AI without providing an opt-out option.
Within Zoom’s legal documentation, users’ video, audio, and chat transcripts are classified as “Customer Content.” By agreeing to the terms, users grant Zoom “perpetual, worldwide, non-exclusive, royalty-free, sublicensed, and transferable license and all other rights” to utilize this content for “machine learning, artificial intelligence, training, testing,” and other product development objectives.
The company also reserves rights for “Service Generated Data,” encompassing telemetry data, product usage data, diagnostic data, and other insights derived from analyzing users’ content and behavior.
Despite the clarification, users remain cautious, calling for clearer opt-out mechanisms and more transparent policies. The incident raises concerns about user privacy and data usage, prompting discussions about the ethical utilization of user data in AI development.
