After Backlash, Zoom Now Says It Won't Train AI Systems On Customer Content

An anonymous reader quotes a report from Variety: Zoom changed its terms of service to say that it won't use any customer content -- at all -- in training generative artificial intelligence models. The update, which the videoconference company announced Friday, comes after observers raised the alarm about a recent change in Zoom's TOS that appeared to grant the company royalty-free rights in perpetuity for customer video calls and presentations for the purposes of training AI models. In its initial response on Aug. 7, Zoom said it doesn't use any customer audio, video or chat content for training AI "without consent." Now it says it will not use such content in any way related to generative AI development.

In a statement Friday appended its its earlier blog post, Zoom said, "Following feedback received regarding Zoom's recently updated terms of service, particularly related to our new generative artificial intelligence features, Zoom has updated our terms of service and the below blog post to make it clear that Zoom does not use any of your audio, video, chat, screen-sharing, attachments or other communications like customer content (such as poll results, whiteboard, and reactions) to train Zoom's or third-party artificial intelligence models." Zoom said it also updated in-product notices to reflect the change. According to Zoom's revised terms of service, the company still owns all rights to what it calls "service-generated data." That comprises telemetry data, product-usage data, diagnostic data and similar data "that Zoom collects or generates in connection with your or your End Users' use of the Services or Software," the terms of service say.

we won't use it..

[laktech]

we'll just sell it to your partners.

Re:

[ls671]

Good, so the set up we planned will still be useful then! At our humorists association, we plan to do a once a month video conference where we will talk very seriously without laughing but only we'll say things that make absolutely no sense at all. Then, we assume the AI will repeat it because apparently it's what pseudo AI (LLM) does.

We may even fork a TV show like "fool the AI" or something like that afterwards. What we will say should still make you laugh although but the goal is that it won't look or s

Re:

[bagofbeans]

Sure. And if they do use it, how do you prove it, even under civil standards (preponderance)?

Re:

[jhoegl]

Doesnt matter if they dont change their TOS back to reflect this change.

This means they say they wont do it, but without changing terms of service, and the "hereby agree to allow us to use AI" part, then they can say "Oh we wont do that", and still do it because you agreed to it in the TOS.

Re:we won't use it..

[postbigbang]

Lots of legal departments are doing hand-wringing over the entire debacle. None of this bodes well for Zoom or its stock price.

Someone made a bad decision, and many are looking for alternatives, ranging from FOSS to (OMG) Teams. It's going to hurt, no matter the direction.

Re:

[ppanon]

Yeah, that's not "a bad decision". That's a "how the heck did this completely moronic idea of using peoples' private conversations, potentially containing corporate secrets, to train AIs to regurgitate similar info make it past approval by multiple layers of management?" How did someone not stand up and say "This is going to blow up in our faces and alienate our customer base"? That's the real reason why people should seriously think twice about ever using their product again - lots of people in the Zoom c

Smell that?

[RitchCraft]

It's the smell of BS coming from a Chinese Zoom kitchen.

I sense a distraction.

[MobileTatsu-NJG]

K... but what about the part where you're demanding a return to the office?

Service-generated data?

[sonamchauhan]

Automatic text captioning of speech in Zoom sessions is service generated. Zoom owns that?