Zoom has recently updated its terms of service due to negative feedback it received for allowing AI training on customer data without proper consent. The changes, implemented in March without much attention, initially granted the company extensive control over customer data for AI training purposes. However, after facing criticism, Zoom clarified its stance through a blog post. The company now states that it will not train AI models on customer video, audio, or chats without obtaining the customer’s explicit consent.
One of the main concerns arose from Zoom’s experimental AI tools, such as IQ Meeting Summary and IQ Team Chat Compose, which use machine learning and artificial intelligence. While account owners must provide consent before using these tools, other participants are only given the choice to accept the terms and join the meeting or decline and leave.
The revised terms address these issues and emphasize that customer data will be used solely to enhance the performance and accuracy of AI services. Zoom’s Chief Product Officer, Smita Hashim, asserted that customer data shared for AI training will not be used for any third-party models. Additionally, Zoom customers will retain ownership and control over their content, even if Zoom provides value-added services using that content, like livestreaming a webinar on YouTube.
The company made it clear that they will not use customer content, including sensitive information like education records or protected health data, to train their AI models without the customer’s consent. The new terms aim to give Zoom account owners and administrators more control over these features and decisions, ensuring transparency and respect for user privacy and consent.
Frequently Asked Questions (FAQs) about AI training
Q: What changes did Zoom make to its terms of service regarding AI training on customer data?
A: Zoom updated its terms of service to require customer consent for AI training on their video, audio, and chats. The changes aim to enhance privacy and give users more control over their data.
Q: What sparked the blowback against Zoom’s policy changes?
A: The blowback was triggered by the discovery that Zoom’s terms allowed them to use customer data for AI training without providing an opt-out option, raising concerns about user privacy and consent.
Q: Which experimental AI tools caused concern among users?
A: Zoom’s experimental AI tools, such as IQ Meeting Summary and IQ Team Chat Compose, which use machine learning, were among the sources of concern due to the data usage policy.
Q: How does Zoom address the concerns related to AI training on customer content?
A: Zoom’s Chief Product Officer clarified that consent is required from account owners and administrators before sharing data for AI training. The data will solely be used to improve AI services and not for third-party models.
Q: Does Zoom retain ownership of customer content?
A: No, Zoom customers retain ownership and control over their content. Even when Zoom provides value-added services using customer content, the original content remains under the customer’s ownership.
Q: What sensitive data does Zoom promise not to use for AI training without consent?
A: Zoom commits not to use customer content, including education records or protected health information, for training their artificial intelligence models without the customer’s explicit consent.