Zoom Updates Terms of Service, Assures Users that Customer Data Will Not be Used to Train AI
In response to widespread criticism and concerns regarding the training of artificial intelligence (AI) using customer data, Zoom has clarified its position on the matter. The company issued an update to its previous announcement, explicitly stating that it will not utilize audio, video, chat, or any similar data to train its AI models.
The clarification comes after Zoom’s Chief Product Officer, Smita Hashim, published a blog post earlier this week, which initially raised concerns about the company’s intentions. The terms of service mentioned that Zoom could potentially use customer content, including audio, video, chat, attachments, and screen-sharing, to train its in-house or third-party AI models. However, in a subsequent blog post on Monday, Hashim assured users that Zoom would not use customer content for AI training, with a few exceptions.
To further solidify its commitment, Zoom has now revised Section 10 of its terms of service, eliminating any legal rights to employ customer content for AI model training. It is important to note that Zoom has not provided specific details about the data sources that will be used to train these AI features. Gizmodo has reached out to the company for further clarification but has not received an immediate response.
Additionally, Zoom has updated Hashim’s blog post to include an apologetic note, acknowledging the feedback received regarding the recent terms of service update. The updated blog post now explicitly states that Zoom will not use any audio, video, chat, screen-sharing, attachments, or other communications, such as poll results, whiteboard content, or reactions, to train its own or third-party AI models.
In conjunction with these developments, Zoom announced new generative AI features that will be introduced to its platform. Zoom IQ Meeting Summary aims to provide users with a comprehensive summary of their meetings, while Zoom IQ Team Chat Compose is a text generation tool. Zoom has assured users that all participants in a meeting will receive on-screen notifications when these AI tools are enabled.
AI training using user data has been a contentious issue, with concerns arising from the increasing prevalence of artificial intelligence applications. OpenAI’s GPT-4, considered one of the most influential AI models, has frequently faced criticism due to its lack of transparency. When GPT-4 was released, OpenAI declined to disclose the data sources used for training, leading to speculation about the quality and origins of the information being generated.
Zoom’s commitment to not utilizing customer data for AI training showcases its determination to address privacy concerns and tread carefully in the realm of artificial intelligence. By clarifying its position and updating its terms of service, Zoom is taking steps to ensure user confidence and demonstrate that it prioritizes data privacy and security.
As AI continues to advance, it is crucial for companies to navigate this landscape responsibly and acknowledge the need for transparency. Zoom’s recent actions reflect a commitment to honoring user trust and maintaining ethical AI practices.
