Live data from Hacker News

Zoom terms now allow training AI on user content with no opt out

explore.zoom.us

121–130 of 538 posts

Re: Zoom terms now allow training AI on user content with no opt out

#121

Earlier quoted context omitted.

It's useful. It's more of a large-scale broadcast situation. Think of large corporate town halls, town council meetings, etc.

You can just have a conference call with the 5-10 speakers and use broadcasting software to stream it to the audience, why do they need to be in the conference?

Why setup a separate broadcast when listeners can just join the meeting room?

Re: Zoom terms now allow training AI on user content with no opt out

#122
post #30
post #14

Earlier quoted context omitted.

Before that: > Customer Content does not include any telemetry data, product usage data, diagnostic data, and similar content or data that Zoom collects or generates in connection with your or your End Users’ use of the Services or Software (“Service Generated Data”). I could be wrong, but my take is that there is not all that much to see here

It's hard to understand what they mean. I understand it as they're free to generate "Service Generated Data" based on “Customer Content”. So for example, a compressed rendition of a call recording would be "Service Generated Data" and thus they will be free to do whatever they want with it (improve their caption generation models ... or sell it to someone?).

> It's hard to understand what they mean.

Working as designed, surely.

Re: Zoom terms now allow training AI on user content with no opt out

#124

Earlier quoted context omitted.

You can just have a conference call with the 5-10 speakers and use broadcasting software to stream it to the audience, why do they need to be in the conference?

Live Q&A is a nice feature.

Conference for the speakers + unlisted livestream on YouTube could handle that, using chat for Q&A.

Re: Zoom terms now allow training AI on user content with no opt out

#125

I am curious if they have been silently saving voice to text transcription in the background on all calls and if AI will be permitted to ingest all of that data. A great deal could be learned from private one on one calls in the corporate world. The insider knowledge one could gain about corporations and governments would be fascinating.

I feel as if 2023 could become the inflection point where we will finally start investing in our own infrastructure again. Video calls for example are really a commodity service to be set up at this point.

Re: Zoom terms now allow training AI on user content with no opt out

#126

Earlier quoted context omitted.

Good catch jxf! but what is that boundary line between SGD and Customer Input/Content? Is it blurry or clearly defined? It seems like things like translations or future enhancements might fall into that area (it also seems like training AI on diags isn't as useful), so this might be expanded in the future now that they have that language in place.

It is defined not at all. Sorry if this is bad for your investment decisions, jxf, but this company is not trustworthy.

I don't have any positions in Zoom (although I did have some puts last year that I've since closed out).

Re: Zoom terms now allow training AI on user content with no opt out

#127
post #40
post #15

What are the alternatives people have tried and had good experiences with?

I have tried most of them: Google Meet, Teams, Slack, Discord, Skype, Jitsi and so far I liked Jitsi the most and Skype the least.

There is also Nextcloud Talk

Re: Zoom terms now allow training AI on user content with no opt out

#128
post #11

Thankfully nothing like this is in Jitsi Meet’s TOS: https://jitsi.org/meet-jit-si-terms-of-service/ It never ceases to amaze me how companies choose the worst software!

+1 for Jitsi. They are awesome, lightweight, and just work with the least hassle. Pretty bad that many nontechnical users are not aware of it compared to Google Meet or Teams.

A lot of us technical users have never heard of it either lol

Re: Zoom terms now allow training AI on user content with no opt out

#129
The Dawn of A.I. Mischief Models

https://slate.com/technology/2022/08/4chan-ai-open-source-tr...

Adversarial machine learning explained: How attackers disrupt AI and ML systems

https://www.csoonline.com/article/573031/adversarial-machine...

How to attack Machine Learning ( Evasion, Poisoning, Inference, Trojans, Backdoors)

https://towardsdatascience.com/how-to-attack-machine-learnin...

AI Security and Adversarial Machine Learning 101

https://towardsdatascience.com/ai-and-ml-security-101-6af802...

The Road to Secure and Trusted AI

https://adversa.ai/report-secure-and-trusted-ai/

Post reply on HN