Zoom's latest terms of service allows use of customer data for AI efforts

Zoom said it wouldn't use 'audio, video or chat Customer Content' for AI training

Zoom can use certain customer data for its artificial intelligence (AI) efforts after making alterations to its terms of service (TOS).

Under the video communications company’s TOS, Zoom receives consent to use their customers’ "service generated data" for many things. The TOS identified "training and tuning of algorithms and models" for AI as one of the permitted uses.

The changes to Zoom’s TOS happened in March, according to recent reports. 

"Zoom customers decide whether to enable generative AI features, and separately whether to share customer content with Zoom for product improvement purposes," a Zoom spokesperson told FOX Business on Monday. "We've updated our terms of service to confirm that we will not use audio, video, or chat customer content to train our artificial intelligence models without your consent." 

ZOOM ASKS EMPLOYEES TO RETURN TO OFFICE FOR FIRST TIME SINCE COVID-19 PANDEMIC

Zoom said in the TOS that "telemetry data, product usage data, diagnostic data, similar content or data that Zoom collects or generates in connection with your or your End Users’ use of the Services or Software" fell under the "service generated data" category.

Zoom Video

In this photo illustration a Zoom Video Communications logo seen displayed on a smartphone. (Photo Illustration by Rafael Henrique/SOPA Images/LightRocket via Getty Images) (Rafael Henrique/SOPA Images/LightRocket via Getty Images / Getty Images)

The document also has a section in which Zoom customers grant the video communications company "perpetual, worldwide, non-exclusive, royalty-free, sublicensable, and transferable license" on customer content for AI, product development and quality assurance, among other things. 

Ticker Security Last Change Change %
ZM ZOOM VIDEO COMMUNICATIONS INC. 78.93 -2.21 -2.72%

Zoom Video Communications Inc

The TOS does state, however, that Zoom "will not use audio, video or chat Customer Content" in the training of its AI models "without your consent," something the company also emphasized in a Monday post from Chief Product Officer Smita Hashim.

ZOOM REDUCING HEADCOUNT BY 1,300

In the post on its website, the company said it updated its TOS for transparency purposes around use and ownership of content on Zoom. For "service generated data," Zoom said it "consider[s] this to be our data" so the company can use it "to make the user experience better for everyone on our platform."

For AI training, Zoom also said "education records or protected health information" were off-limits without consent.

Over shoulder view of female worker have webcam digital virtual conference with diverse multiethnic colleagues. Downdetector, which reports service outages, said reports of issues with the digital video conference service skyrocketed within the past (istock / iStock)

The company said users have the option to give consent to AI model training when two of its recently-launched AI features get turned on. In those instances, the user’s data only goes toward boosting the "performance and accuracy" of the features and not for third-party models, according to Zoom. 

The AI-powered Zoom IQ Meeting Summary and Team Chat Compose features have been available for roughly two months.

Zoom building

Zoom building (Getty)

Zoom has previously said it has its own AI models, in addition to utilizing ones from companies like OpenAI and Anthropic and those of certain customers. OpenAI is the Microsoft-backed firm behind ChatGPT, an AI chatbot that has garnered significant attention and popularity since its debut late last year.

AI HAS ALREADY SURPASSED HUMAN CAPABILITIES IN SEVERAL DOMAINS; WHICH JOBS ARE NEXT?

The value of Zoom’s stock has experienced a roughly 3.3% increase from where it hovered at the beginning of the year. In the past 12 months, it has fallen almost 40%.