Zoom has made changes to its terms of service in response to an online post about recent policy updates that allow its artificial intelligence to be trained using customer data. A StackDiary article last weekend highlighted how the changes, introduced in March, appeared to give the company a comprehensive overview of its customer data for AI training purposes.
See also: Zoom calls coming soon to Sony Bravia TVs
In response, Zoom published a blog post claiming that it would not enforce its terms. The company then updated its terms in response to the ongoing debate on the issue, stating that it will not train AI using video, audio, or consumer conversations without the customer’s consent.
At least part of the problem stems from Zoom’s experimental AI tools, which include Intelligent Meeting Support IQ and Team Communication IQ Chat (AI-powered message composition). Although users must give consent before starting a session using these tools, additional participants have only two options: accept the terms and join the session, or reject them and leave the session.
“ What’s troubling is the company’s explicit mention of its right to use this data for machine learning and artificial intelligence, including training and tuning algorithms and models ,” Alex Ivanovs wrote for Stack Diary. “ This essentially allows Zoom to train its AI on customer content without providing an opt-out option, a decision that is likely to spark significant debate about privacy and consent user .”
See also: Europe: Zoom offers local data storage to premium customers
Ivanovs emphasized that his terms grant the right to “redistribute, publish, import, access, use, store, transmit, revise, disclose, preserve, export, modify, reproduce, share, use, display, copy, distribute, translate, transcribe, create derivative works from, and perform any and all acts with respect to the Customer Content.”
In a company blog post published on Monday, Zoom Chief Product Officer Smita Hashimstressed the importance of obtaining consent from owners and administrators before deciding to share their data for AI training. She insists that this data is used solely to improve the performance and accuracy of AI services. Hashim added that “even if you choose to share your data, it will not be used to train third-party models.” She continued: “We have permission to use this customer content to provide services based on this content, but our customers continue to own and control their content. For example, a customer may have a webinar that they ask us to live stream on YouTube. Even if we use the customer’s video and audio content for live streaming, they own the underlying content.”
“We will not use customer content, including training records or protected health information, to train our AI models without your consent,” the blog post says. A new section added to Zoom’s terms makes that clearer: “Despite the above, Zoom will not use audio, video, or chat content to train our AI models without your consent.”
See also: OpenAI's artificial intelligence is coming to Zoom in April
“Our goal is to allow Zoom account owners and administrators to have control over these features and decisions, and we’re here to shed light on how we do this and how this impacts certain customer groups,” Hashim wrote.
Zoom, known for providing digital tools for meetings and collaboration, has made significant inroads into the AI market. However, the company’s recent policy change regarding training AI models based on customer data has raised concerns about data protection. With AI becoming increasingly critical in the digital age, changing privacy standards and the resulting consent issues must be addressed with transparency and accountability.
