Hi there, I'm the COO of Zoom. We currently do not use audio, video or chat content to train AI models and we would not do so without customer consent. Please see our blog https://blog.zoom.us/zooms-term-service-ai/
> We do not use audio, video or chat content to train AI models
<Edit: the word “currently” was edited in to the parent comment after I originally quoted it>
What your linked article says:
> When you choose to enable Zoom IQ … you will also be presented with a transparent consent process for training our AI models using your customer content.
Consent or not, these two statements seem to be at odds with each other?
There isn’t room for such ambiguity in this discussion. Details really matter here.
This community has a long memory about previous Zoom security issues and misleading statements (eg about encryption) - so it’s worth understanding that the level of scrutiny being given to this latest topic is likely elevated as a result. (In other words, I think Zoom has some extra work to do to rebuild trust among many of us here).
Hi there, I don't believe you, especially since you've misrepresented that blog post. It actually says: "we do not use audio, video, or chat content for training our models without customer consent." It explicitly says that you do use customer content to train your models.
From the screenshot: "You are directing Zoom to access, process, and use your participants' inputs and AI-generated content for the purpose of Zoom IQ product improvement, including model training."
Are you claiming that agreeing to these terms of service doesn't count as consent? I think your lawyers would disagree.
As the others have been saying, you need to absolutely spell this out in the ToS. The language in the ToS is far too broad and no matter how trustworthy your company may be now, people are cluing into the fact that that trustworthiness is one heartbeat away from a change in leadership that decides to push the boundary up to the legal limits in the law or ToS.
They explicitly said “we do not”, not “we will not.” The ToS and their statements allow them to change their mind at any time without notifying anyone.
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. https://blog.zoom.us/zooms-term-service-ai/
I do not think that blog posts have any legal value whatsoever. The current terms of service are too risky for most companies who would otherwise have loved to continue to be users of your product.
I still remember when they promised E2E encryption when they had at least a MITM on their end. They have since deleted their (whole!) blog where the whole bullshit non-apology was posted. Zoom’s comments or promises are completely worthless.
“Customer consent” as in, accepting the TOS by continuing to use the product? Have some respect for the intelligence of your users. If you’re playing dirty lawyerball in HN comments you can’t be trusted to act in good faith.
Say more things. Are you suggesting that by “customer consent”, they mean the consent of someone other than those of us paying for zoom? That makes no sense, and is not supported by the use of “customer” in their TOS.
No. I'm saying "you" are the person in the Zoom call and who agreed to the Terms of Service when you installed the software, but the "customer" is the Zoom account owner, and these are usually not the same person.
Most people in Zoom calls at any moment are in calls where they are not the customer for that call. I would guess most Zoom users have never been customers in that sense, as in they never initiate calls and may not have an account.
Sure, and I think all-party consent should be required if a call has any guests. My point was much shallower—just that OP/Zoom is using language that suggests “customer consent” is some separate thing they would ask for, when in fact using the software (accepting the TOS) is that consent.
Thank you for your feedback. 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. https://blog.zoom.us/zooms-term-service-ai/
It's especially annoying because I can count on both hands situations where the blogging / social media team and the engineering / legal teams weren't on the same page. The blogging team sends out a message saying "we don't," while engineering is already working on something that does and legal has signed off because it doesn't violate the Terms.
She's lying, if you click on the blog post you can see that if an account owner opts in, they use audio, video, and chat content of all participants to train their models. As a user you are notified of this but not given a way to opt out, other than just leaving the meeting.
Based on feedback, 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.https://blog.zoom.us/zooms-term-service-ai/
That's nice, but your comment still says "We currently do not use audio, video or chat content to train AI models". You do, currently. If you don't think there's anything wrong with that, why come here to lie about it?
"without your consent" is also very misleading - you get the account owner's consent, but if I join another company's Zoom meeting and they have this turned on, I am not meaningfully asked for consent. I am informed, but the only way to "opt out" is to leave the meeting immediately (but you already have my face?) and uninstall Zoom. This is obvious coercion, Zoom has majority market share and most people do not have the freedom to decline when somebody sends them a meeting link.
"We don't use your data to train AI models" (but we can pass it to another entity to do so)
"we would not do so without customer consent" (...and everyone who uses our service has consented to anything we want them to in our convoluted ToS.
What you should have said is: "Zoom doesn't use audio, video or chat content to do anything but transmit it to the parties that are logged in to the meeting."
I frequently have to use Zoom for communications that are privileged by law. I wonder when the class action will happen?
Based on feedback from you all, 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. https://blog.zoom.us/zooms-term-service-ai/
There's definitely more that can be done in terms of setting trust with your users, partners, employers, and suppliers - I posted a comment elsewhere on whether or not you have a posted and shared Trustworthy AI policy, and something that is used to guide internal AI development. Happy to email that to you as well.
Even if you don't, the language of the agreement allows you to. The very same clause contains the consent language, so now that you have the materials, you can start making fake versions of us at your leisure.
Hi COO, of zoom, when people leave or join our zoom meetings the CHIME does not work when the host has left audio. this makes our business really hard to operate. as we rely on your awesome product. HALP
