In response to user concerns and scrutiny over its data usage policies, video-calling giant Zoom has recently clarified its stance on utilizing user data for artificial intelligence (AI) training. The company’s move comes after a revision of its terms of service earlier this year sparked fears that it could potentially exploit audio, video, and chat data for AI purposes without user consent.
The firm’s initial alterations to its terms of service in March prompted a wave of unease among users who believed these changes granted the company broader access to their personal data for AI training. Data protection experts had expressed concern about the vagueness of the terms, which appeared to permit Zoom to employ user-generated content for diverse purposes.
Leading up to the update, data protection specialist Robert Bateman highlighted that the original wording of the terms of service presented significant concerns. Bateman pointed out that such sweeping contractual provisions could be risky and should raise alarms, urging for more transparency and specificity.
Addressing these apprehensions, Zoom announced a revised version of its terms, specifically stating, “Zoom will not use audio, video, or chat customer content to train our artificial intelligence models without your consent.” This move, the company emphasized, aimed to provide users with the assurance that their data would not be leveraged for AI purposes without explicit permission.
In June, Zoom unveiled a series of AI-powered features, including the ability for users to summarize meetings without recording the entire session. These features were introduced as part of a free trial. However, some experts cautioned that the initial terms of service modifications could have potentially allowed Zoom to access more user data than required, including information from customer calls.
While the utilization of AI to replicate human-like behaviors has gained popularity across various industries, it has also sparked concerns related to privacy, sensitive data, and copyright infringement. The Open Rights Group, an organization advocating for digital privacy, expressed concerns that Zoom’s decision to offer AI features as a free trial and encourage customers to “opt in” might obscure the extent of data usage for AI training.
In response, a Zoom spokesperson reiterated that users retained the discretion to enable generative AI features and choose whether to share customer content with Zoom for “product improvement purposes.” The company highlighted its commitment to transparency by showcasing warning messages for users participating in meetings that utilize AI tools. These messages grant users the option to either consent to AI training usage or exit the meeting.
Smita Hashim, Zoom’s chief product officer, emphasized that account owners and administrators had the power to activate the AI features, which were still being offered on a trial basis. Hashim assured that individuals who opted for these features would be guided through a transparent consent process, assuring them that their customer content would be used for AI model training only with their explicit agreement.