Zoom has revised its terms of service in response to feedback about its plans to use customer data for training its artificial intelligence (AI) and machine-learning models. The video conferencing platform recently updated its terms of service, which allowed Zoom to access, use and modify customer data for the purpose of training and refining its AI and machine-learning models. Additionally, customers were also required to grant Zoom a perpetual, worldwide, non-exclusive, royalty-free license to use their data for product and service development as well as for AI and machine learning.
However, Zoom has recently made further updates, including a new clause, which explicitly states that it will not utilize audio, video, or chat customer content for the purpose of training its AI models without obtaining the user’s consent. The official blog post states, “We have revised our terms of service (in section 10.4) to provide additional assurance that we will refrain from using audio, video, or chat customer content to train our artificial intelligence models unless we have obtained your consent.” “
The changes were implemented in response to a recent company update, where it altered its policies and revised its terms and conditions, granting the company a significant level of control over customer data for the purpose of AI training, as highlighted in a report by StackDiary.
Pursuant to Section 10.2 of Zoom’s Terms of Service, Customers provide their consent for Zoom to access, use and manipulate data generated through the Service, including its application in machine learning and AI efforts. Furthermore, section 10.4 of Zoom’s terms and conditions states that users “grant Zoom a perpetual, worldwide, non-exclusive, royalty-free, sublicensable license to employ their data for the development of products and services and agree to grant a transferable license”, as well as for the advancement of machine learning and artificial intelligence.
The modifications sparked a backlash among Zoom’s user community due to concerns about user privacy, prompting the company to amend its policy. Meanwhile, Zoom’s Chief Product Officer Smita Hashim issued a clarification via a blog post stating that Zoom users must provide their consent before their data is shared for AI training, “specifically for these AI employed to enhance the performance and accuracy of the services.”
He further explained that “even if you (the user) choose to share your data, it will not be used for training any third party models.”
“We are permitted to use this customer content to provide value-added services based on such content, yet our customers retain ownership and control of their own content. For example, a client may request us to livestream a webinar. Even though we use their video and audio content for livestreaming, Hashim clarified, “The original content remains their property.”
In addition, a new section has been introduced in Zoom’s Terms of Service to provide clarity: “Notwithstanding the above, Zoom may not use audio, video or customer chats for the purpose of training our artificial intelligence models without your explicit consent.”
However, even with this change, Zoom’s privacy policy doesn’t clarify what “service generated data” actually means, which can be confusing. The policy still allows Zoom to use customer information for various things like teaching AI, and it doesn’t seem possible to say no in Zoom’s settings.
Meanwhile, Zoom’s advances in AI are in line with the ongoing transformation in the tech industry driven by artificial intelligence. Companies like Google, Microsoft, Meta, Snapchat and even Truecaller are currently starting the AI journey. Zoom, in particular, has been actively developing its AI models to enhance the convenience of meetings and other platform features.
