Zoom Won’t Use Your Data to Train AI. Period.

Written By :

Category :




Posted On :

Share This :

Zoom got themselves in hot water a couple weeks ago for not clearly communicating about use of your data to train their AI models.


Amongst the legitimate concerns was a lot of misinformation (and maybe even hysteria) that Zoom was using your calls without your consent to train their systems.  They weren’t.

Nevertheless Zoom responded to that concern by confirming that Notwithstanding the rest of TOU section 10.4 “Zoom will not use audio, video or chat Customer Content to train our artificial intelligence models without your consent”.  I walked through the concerns and impacts in a recent blog post here.

Now, whether from pressure from their clients (many of the largest and most security and privacy reliant organizations in the world!), or just out of an abundance of desire to restore and maintain trust, they have changed their practice (and their TOU) to not use customer data for training AI at all.  Not third party AI, not their own AI.  Not even to use your data to make the AI more effective to serve you apparently.  Not even with your consent.

Zoom updated TOU removing "Without Consent" regarding training AI
Zoom removed the “without your consent” wording as they will no longer use your data for this purpose period.

Actually you can’t even give your consent if you want to anymore as they took that option away! 


So even if you fully trust Zoom as many of the world’s leading experts and companies do, and you’d like to use your data to help perfect the AIs for your benefit, you cant.  There’s no long an option for that.  They removed the option and replaced it with a confirmation that “Zoom does not use any 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 AI models” as seen below.


ZoomIQ removes "without consent" from privacy policy
Zoom used to provide an option to OPT-IN to share data with Zoom to train the AI. Now it’s gone.

I’m not sure how smart that is from an improving AI to serve us better perspective (something I think we all want), but as Eric Yuan (Zoom’s CEO) knows, trust is crucial and since trust was diminished (not because they did anything majorly wrong necessarily but more in terms of mistakes they made in how they communicated about it), trust must be restored.  

As such, no training of AI with customer data, period.


You can find Zoom’s updated TOU agreement here.

How do you feel about this?  Was this the best move?  Why?