Is Zoom training AI with user data?
With artificial intelligence (AI) still on the up and up, tech companies are looking into how they can best use AI tools in their own products and services.
A few minutes every morning is all you need.
Stay up to date on the world's Headlines and Human Stories. It's fun, it's factual, it's fluff-free.
The backstory: With artificial intelligence (AI) still on the up and up, tech companies are looking into how they can best use AI tools in their own products and services. The thing is, generative AI needs to be trained on relevant data in order to be useful in these contexts. AI farms already-existing content to help humans solve problems in a streamlined way, but it can’t necessarily “think” for itself, instead combing through stuff that’s already out there. And a lot of people don’t want their data being repackaged and used in these ways without consent. We’ve seen this kind of thing emerge as a problem among artists and other creators – but what about non-creatives whose info is all over the net?
More recently: In March, video-conferencing software company Zoom made some changes to its terms of service (TOS) that raised some eyebrows with its user base. The wording of one clause seemingly gave Zoom general permission to use customer data to train its AI products. Around this time, experts started warning people that, according to Zoom’s TOS, the company might be able to pull from a lot more user data than expected, like via user calls. In June, Zoom launched new AI features, with one helping users summarize meetings without a full recording of the call – these features are currently available on a trial basis. These developments brought its TOS under the microscope again.
The development: Almost right after these new terms became publicly available, Zoom started seeing backlash. On Monday, the company came out with a blog post explaining that it wouldn’t use audio, video and chats to train AI without customer consent. Soon after, the Zoom team re-updated the TOS to make it clear that no audio, video or text chats would be used for AI training without explicit consent.
Now, attached to the clause that gives Zoom permission to collect “Service Generated Data” and allows it to use that data for “product and service development,” there’s a new statement that promises not to use “Customer Content” without consent. But, Zoom isn’t the only tech company that collects user data or that includes AI-powered features, so it’s safe to say people may keep a closer eye on these TOS details.
Key comments:
"The terms appeared to give the service provider a lot of freedom to use data generated by its users for many different purposes,” data protection specialist Robert Bateman told the BBC.
It’s important to us at Zoom to empower our customers with innovative and secure communication solutions,” Zoom’s Chief Product Officer, Smita Hashim, wrote in a blog post on Monday. “We’ve updated our terms of service (in section 10.4) to further confirm that we will not use audio, video, or chat customer content to train our artificial intelligence models without your consent.”
"Zoom will not use audio, video or chat customer content to train our artificial intelligence models without your consent,” Zoom’s new TOS read, updated Monday, August 7.
Comments ()