Zoom, the popular video conferencing platform, has recently updated its terms of service. The move sparked discussions about the extent to which companies can use customer data to train AI. The update, effective since July 27, grants Zoom the right to leverage certain aspects of customer data to train and fine-tune its AI and machine-learning models. This move showcases Zoom’s ambition to bolster its AI capabilities for improved user experiences.
What Customer Data Will Zoom Use for AI Training
The updated terms of service mention that Zoom can utilize “service-generated data” to train its AI models. The category includes valuable customer information. These include product usage data, telemetry, diagnostic data, and other relevant content collected by the company. It’s worth noting that this data usage category doesn’t offer an opt-out option. It appears that Zoom intends to harness a diverse range of data sources for AI enhancement.
Zoom Issues Clarity on its Data Usage
In light of growing concerns regarding data privacy and AI training practices, Zoom has clarified its stance. The company emphasized that customer consent is crucial to its data utilization approach. Zoom does not use audio, video, or chat content for training its models without explicit customer consent. This clarification addresses any potential concerns about the misuse of sensitive information.
Generative AI Features and Consent
In June, Zoom introduced new generative AI features, including a meeting summary and chat message composition tools. These features are available on a free trial basis for customers who can choose whether or not to use them. However, when users opt for these features, they must sign a consent form allowing Zoom to utilize their individual customer content for training its AI models. The consent form highlights that the data collected is solely intended to enhance the performance and accuracy of the AI services.
Industry Context and Implications
Zoom’s updated terms of service come amidst a broader debate about the ethical and legal aspects of training AI models using personal or aggregated data. Similar concerns have arisen in the generative AI sector, with some individuals claiming that their work is being replicated in AI-generated outputs. This situation raises questions about intellectual property rights and the boundaries of AI training practices.
Zoom’s Response and Customer Empowerment
A spokesperson for Zoom emphasized that the company values customer choice and control. Zoom users have the autonomy to enable generative AI features and decide whether to share their customer content with the company for product improvement. This approach underlines Zoom’s commitment to putting the power of data usage decisions into the hands of its users.