Zoom has updated its terms of service to indicate its intention to enhance its artificial intelligence capabilities through the utilization of certain customer data. The updated terms, effective from July 27, emphasize Zoom’s authorization to incorporate specific aspects of customer data in the training and fine-tuning of its AI and machine-learning models.
The newly permitted “service-generated data” encompasses data related to customer usage patterns, telemetry, diagnostic information, and comparable content collected by Zoom. Notably, the updated terms do not offer an opt-out mechanism.
While the practice of employing such data for AI training is not uncommon, these modified terms signify a measured progression toward Zoom’s own ambitions in the field of AI.
This alteration arrives in the midst of a broader public discourse concerning the extent to which AI systems should be trained using individuals’ data, regardless of the claimed aggregation or anonymization. Numerous AI systems, including chatbots like OpenAI’s ChatGPT, Google’s Bard, and Microsoft’s Bing, as well as image generation tools, draw from extensive portions of internet text and imagery. In recent times, legal actions have arisen from authors and artists who assert that AI outputs mirror their original creations.
Zoom’s terms explicitly state that users “consent to Zoom’s access, use, collection, creation, modification, distribution, processing, sharing, maintenance, and storage of Service Generated Data for any purpose, to the extent and in the manner permitted under applicable Law, including for the purpose of… machine learning or artificial intelligence (including for the purposes of training and tuning of algorithms and models).”
Notably, customer content like messages, files, and documents appears to be excluded from this category. In a subsequent blog post, Zoom clarifies that they do not employ audio, video, or chat content for training their AI models without customer consent. The crucial phrase here is “without customer consent.”
Earlier in June, Zoom introduced two new generative AI features, namely a meeting summary tool and a chat message composition tool, available on a trial basis for customers. Users who opt to utilize these features are required to provide consent for Zoom to train its AI models using their specific customer content.
Zoom assures users that their content will be employed solely to enhance the effectiveness and accuracy of these AI services. A spokesperson from the company emphasized that it is ultimately up to Zoom’s customers to decide whether to activate generative AI features and whether to share customer content for the purpose of product improvement.