Zoom has just updated its terms of service to reflect it can now use certain aspects of customer data to train and fine-tune its AI and machine learning models (via CNBC).
According to the new terms, effective from July 27, Zoom can utilize customer product usage, telemetry, diagnostic data, and similar content for AI training, termed “service-generated data.”
Unlike some other companies, Zoom does not provide an opt-out option for customers in this regard.
This development comes amidst a broader public debate on the ethical use of individual data for training AI services, even if the data is aggregated or anonymized.
Many AI services, including chatbots and image-generation tools, rely on vast amounts of internet text or images for training, leading to lawsuits from authors and artists who claim to see their own work reflected in AI outputs.
Zoom’s updated terms clearly state that customers consent to the company’s access, use, collection, modification, and storage of service-generated data for various purposes, including AI training.
However, the terms seem to exclude customer content such as messages, files, and documents from being used in this way.
Zoom clarified in a follow-up blog post that it does not use audio, video, or chat content for training AI models without customer consent.
For the new generative AI features introduced in June, such as the meeting summary tool and chat message composer, customers are given the choice to enable or disable them.
If enabled, customers must sign a consent form allowing Zoom to use their individual customer content for training its AI models.
A Zoom spokesperson emphasized that customers have the final say in enabling generative AI features and sharing customer content with Zoom for product improvement purposes.
