` `

Zoom Clarifies AI Training Policy, Denies Call Usage Without Consent

Wesam Abo Marq Wesam Abo Marq
Technology
13th August 2023
Zoom Clarifies AI Training Policy, Denies Call Usage Without Consent
Zoom emphasized the importance of users' consent (Getty)

The video conferencing and messaging application Zoom faced criticism and user apprehensions following a modification to its terms of service. This alteration appeared to give Zoom the capability to utilize user-generated "content" for the purpose of refining artificial intelligence (AI) algorithms. 

The company emphasized that it does not employ audio, video, and chat data for AI purposes without obtaining consent from users.

Zoom Company Faced Backlash and Concerns

The update to Zoom's terms of service has triggered concerns on social media. Section 10.4 of the Zoom TOS stated that the platform possessed the authority to utilize customer content for various purposes, including product development, marketing, analytics, machine learning, and AI training

This caused alarm among users, prompting them to engage in discussions on social media platforms regarding the possibility of prohibiting Zoom's use within companies and canceling memberships.

One widely shared tweet conveyed, "Zoom terms of service now require you to allow AI to train on ALL your data — audio, facial recognition, private conversations — unconditionally and irrevocably, with no opt out. Don’t try to negotiate with our new overlords."

Another Twitter user expressed outrage, urging fellow users to remove or uninstall Zoom from their devices.

A supporting image within the article body
Photo Description: A screenshot of a tweet about the issue.

Zoom’s Clarification in Response to the Outcry

In reaction to the public outcry, Zoom provided clarity through a blog post, affirming that it refrains from utilizing audio, video, or chat data to train AI models without customer consent.

A supporting image within the article body
Photo Description: A screenshot from Zoom’s blog post.

The company also outlined its AI services, known as Zoom IQ Meeting Summary and Zoom IQ Team Chat Compose, which are designed to create meeting summaries and compose chat messages. These features are accessible on a trial basis at no cost and require account owners to activate them. 

A supporting image within the article body
Photo Description: A screenshot from Zoom’s blog post.

Zoom highlighted its commitment to transparency by ensuring a clear consent process for customers who opt to allow their data to be used for AI model training. This consent is aimed at enhancing the performance and accuracy of the mentioned AI services, with the assurance that the data will be used solely for this purpose.#

Experts: Concerns About the Use of AI Will Continue

While this event ignited public discourse, experts predict that apprehensions regarding AI usage will persist. Across various industries, companies are revising their terms of service to accommodate novel products and services involving AI. These terms are designed to safeguard the interests of the company, but they must harmonize with privacy policies that elucidate the commitments made to users concerning their personal information.

The utilization of customer content for AI model training gives rise to queries concerning consent and the boundaries of publicly accessible data. Some contend that consent should be obtained before utilizing individual data pieces for model training. In the context of Zoom, decisions about AI usage during calls are entrusted to the account holder. Users who harbor concerns about participating in AI training can request the host withdraw consent or choose to exit the call.

This occurrence underscores the ongoing debates encircling AI use and underscores the significance of lucid terms of service and privacy policies to address user concerns, safeguard the company's interests, and uphold the privacy of individuals.

Human-Like Behavior

Human-like behavior refers to actions, reactions, and patterns of interaction exhibited by artificial intelligence applications that closely resemble those performed by humans. These behaviors are learned and replicated through extensive data analysis and algorithmic processes, allowing AI systems to simulate and imitate human responses in various contexts.

A supporting image within the article body
Photo Description: A screenshot of an MIT News article.

However, the widespread collection of online information for training the foundational models of AI applications has instilled concerns and triggered legal actions due to the potential inclusion of personal, sensitive, or copyrighted content within their datasets.

Zoom, like numerous other technology companies, has increased its emphasis on AI products in response to the escalating enthusiasm surrounding this technology.

Nevertheless, the Open Rights Group, an advocate for digital privacy, has cautioned that Zoom's choice to introduce these features as part of a free trial and to encourage customers to willingly participate in them has made the alterations even more disconcerting.

Read More

Teachers Can’t Hear Muted Students on Zoom

Man Kills Father During Zoom Meeting

Misbar’s Sources:

Zoom

BBC

AP

MIT News