Zoom Clarifies New Terms of Service Amid Concerns User Content Is Used to Train AI

Small toy figures are seen in front of a Zoom logo in this illustration picture taken on March 15, 2021. (Dado Ruvic/Reuters)

Videoconferencing platform Zoom has denied reports that it is using customer data to train artificial intelligence without users’ consent after the company’s updated Terms of Service sparked concern.

In an Aug. 7 blog post, the company sought to clarify the changes and explained it will not use user data to train AI models unless customers explicitly provide their consent.

The San Jose, California-based company initially updated its terms of service in March “to be more transparent about how we use and who owns the various forms of content across our platform,” according to its website.

Under Section 10.2 of those terms, Zoom stated that users agree “that Zoom compiles and may compile Service Generated Data based on Customer Content and use of the Services and Software” and that users “consent to Zoom’s access, use, collection, creation” and more of service-generated data “for any purpose, to the extent and in the manner permitted under applicable Law, including for the purpose of product and service development, marketing, analytics, quality assurance, machine learning or artificial intelligence, training, testing,” among others.

Section 10.4 of the terms, meanwhile, stated that users agree to “grant Zoom a perpetual, worldwide, non-exclusive, royalty-free, sublicensable, and transferable license” to “redistribute, publish, import, access, use, store, transmit, review, disclose, preserve, extract, modify, reproduce, share, use, display, copy, distribute, translate, transcribe, create derivative works, and process customer content and to perform all acts with respect to the customer content.”

The section further stated that this would again be used for the purposes of product and service development, machine learning, artificial intelligence, and training.

A Zoom App logo is displayed on a smartphone. (Olivier Douliery/AFP via Getty Images)

‘Significant Invasion of User Privacy’

The update has sparked concern among many experts amid an ongoing debate regarding AI safety.

An article published on Aug. 6 on Stack Diary, a tech and development publication, noted that the updated terms could potentially see users’ private video and audio calls or text chats used to train AI, noting that critics were calling it a “significant invasion of user privacy.”

“This effectively allows Zoom to train its AI on customer content without providing an opt-out option, a decision that is likely to spark significant debate about user privacy and consent,” the article stated.

Amid growing concern, Zoom Chief Product Officer Smita Hashim stated in the Aug. 7 blog post that the videoconferencing platform “will not use audio, video, or chat customer content to train our artificial intelligence models without your consent.”

Ms. Hashim noted that the company can use service-generated data to improve the user experience but stressed that meeting recordings are still “owned by the customer” and that Zoom has “a license to that content in order to deliver the service of recording.”

“An example of a machine learning service for which we need license and usage rights is our automated scanning of webinar invites / reminders to make sure that we aren’t unwittingly being used to spam or defraud participants,” she wrote. “The customer owns the underlying webinar invite, and we are licensed to provide the service on top of that content. For AI, we do not use audio, video, or chat content for training our models without customer consent.”

Tesla Inc. CEO Elon Musk attends the World Artificial Intelligence Conference (WAIC) in Shanghai, on Aug. 29, 2019. (Aly Song/Reuters)

Concerns Over AI Dangers

Like many companies, Zoom has been incorporating AI into its user features, recently launching Zoom IQ Meeting Summary and Zoom IQ Team Chat Compose.

The first feature allows meeting hosts to initiate an “AI-generated summary of their meeting powered by Zoom’s own large language models,” while the latter “leverages OpenAI’s technology to help you write messages by using context from your conversation.”

“When you choose to enable Zoom IQ Meeting Summary or Zoom IQ Team Chat Compose, you will also be presented with a transparent consent process for training our AI models using your customer content,” Ms. Hashim stated in the blog post.

“Your content is used solely to improve the performance and accuracy of these AI services. And even if you chose to share your data, it will not be used for training of any third-party models,” she added.

Separately, a company spokesperson also told CNBC in a statement, “Zoom customers decide whether to enable generative AI features, and separately whether to share customer content with Zoom for product improvement purposes.”

Republican lawmakers and other experts including Elon Musk have repeatedly called for strong regulatory control over advanced AI amid growing concerns over its potential dangers.

More recently, a team of researchers from Cornell University raised concerns after they were able to train a deep-learning model to record and steal users’ data simply by listening to their keystrokes with a 95 percent accuracy rate.

The researchers also tested the deep-learning model on Zoom with a 93 percent success rate.

A spokesperson for the company told The Epoch Times in an emailed statement, “Zoom takes the privacy and security of our users seriously.”

“In addition to the mitigation techniques suggested by the researchers, Zoom users can also configure our background noise suppression feature to a higher setting, mute their microphone by default when joining a meeting, and mute their microphone when typing during a meeting to help keep their information more secure,” the spokesperson said.