Zoom has issued a clarification on elements of its terms of service related to the use of user data for AI training after a series of LinkedIn posts criticized the company for allegedly giving users no way to unsubscribe.
Users had objected to several clauses stipulating that Zoom could use customer content, such as uploaded files and data or transcripts and analyzes resulting from Zoom calls, to train and tune algorithms and models for artificial intelligence (AI) and machine learning (ML).
Points 10.2 and 10.4 of the Terms of Service raised particular concerns, as they indicated that Zoom might use the content for the purposes of “machine learning, artificial intelligence, training, testing, service improvement, software or other Zoom products, services.” , and software, or any combination thereof.”
Zoom rejected criticism of its terms and reiterated that customers have a choice over how their data is used.
“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 said. ITPro.
Alongside its popular video conferencing software, the company launched Zoom IQ, an AI assistant capable of summarizing meetings and conference call actions.
RELATED RESOURCE
Unified endpoint management and security in a work-from-anywhere world
Understand what influences security strategies today.
At companies that adopt Zoom IQ, administrators control whether information is shared with Zoom for training purposes and can revoke the consent they may grant to Zoom for these purposes at any future date.
Users can continue to use Zoom IQ’s generative AI features whether or not they opt-in to data sharing.
Zoom also collects “service-generated data” (SGD), the term it uses for diagnostic and telemetry data, and any other data that Zoom collects or generates as a result of use by customers of its services and software.
The terms of the agreement state that Zoom users grant the company a “perpetual, worldwide, non-exclusive, royalty-free, sublicensable, transferable license” to manage SGD as l The company deems it appropriate within the limits it has set.
Greg Wilson, senior software engineering manager at Deep Genomics, wrote a job on LinkedIn urging companies to abandon Zoom as a platform while the practice remains in place.
Aparna Bawa, COO of Zoom replied to the message, explaining that the terms were intended to improve transparency rather than raise concerns.
Furthermore, the company published a blog post in which he addressed the new conditions of service, which came into force in March 2023.
“In section 10.4, our intention was to ensure that if we provided value-added services, such as a meeting recording, we would have the ability to do so without usage rights questions,” wrote Smita Hashim , Director of Product at Zoom. .
“An example of a machine learning service for which we require licensing and usage rights is our automated analysis of webinar invitations/reminders to ensure we are not unintentionally being used to spam or defraud attendees .
“The customer owns the underlying webinar invitation and we are authorized to provide the service in addition to this content. For AI, we do not use audio, video, or chat content to train our models without customer consent.
Some LinkedIn users compared the changes to Google’s recent initiatives.
In July, the search giant changed its terms of service to explicitly allow the company to train its AI models on publicly available data.
A passage in its privacy policy, which referred to the use of publicly available information, such as text from open-access websites to train Google’s language models, was changed to include a reference to products of Google AI such as Bard.
Google specifically said that businesses with business information on a website could index it for use in Google services.
The move sparked criticism from some in the industry, who pointed out that it could give the company an unfair advantage over competitors. It has also been compared to OpenAI, which has reportedly used a large amount of public data to train models such as GPT-4.
OpenAI GPTBot configuration released quietly on August 7, which will allow administrators to prevent their web data from being scraped for inclusion in future models trained by the company.
Other companies face similar questions as the race to AI heats up, with productivity tools like Google Duet AI and Microsoft 365 Copilot facing scrutiny.
In an article on Mastodon, security researcher Kevin Beaumont suggested that changes such as automatically generating transcripts with generative AI could give individuals insight into private corporate meetings via GDPR mechanisms.
“My favorite inconsiderate vector with AI is probably that MS Copilot plans to do meeting summaries and transcriptions… so if you want to know what a company says about you in meetings, wait a year and enter this request for access regarding the GDPR,” Beaumont wrote.
ITPro contacted Google for more information.
