This shouldn’t even be an option, but if you’re gonna allow users to restrict access this way, they should be required to include a privacy policy or otherwise state what they intend to do with our information.
As it is, we’re being asked to hand over our name and email address to third parties with no explanation as to why.
Scratch that, repo owners should not be allowed to harvest personal information period.
What possible reason could there be to allow this? It’s completely unreasonable.
(apparently I can’t edit the original post anymore)
This must be where our reasonable concerns go to die.
I understand the need for a gated model. It will be necessary for those who are in the business. Individuals and communities who want clearer feedback would also have an incentive to use it.
So why would they ask for an email address?
Maybe it’s because that’s the default setting for HF’s gated model, but I still don’t understand the deeper meaning either.
BTW, HF has been looking for requests for the site for about a month now.
Hi @missionfloyd
I understand that your concern stems from some AI models having a gated repo which require the user to input their info.
In short this is a measure to make sure that people are complying with the licence agreement as defined in these repos.
Think of it as you signing a document.
As for removing gated access I have no say in this but I don’t think that is going to happen, the reason behind it is that they work as user complience agreement, even meta and google have gated repos as well (examples : meta-llama/Llama-3.2-11B-Vision-Instruct · Hugging Face , google/gemma-2-2b-it · Hugging Face , etc … ), note that not all AI models have gated access so feel free to switch to another model if you want to.
This isn’t about merely clicking “Agree” on yet another EULA, it’s that they are given access to our personal information. Under no circumstances should any user be allowed to harvest information from other users.
And the fact that google and facebook do it, sort of proves my point. These companies’ entire business models are based on collecting such information for advertising purposes.
Well, that Google and Facebook also offer their AI models in HF, and if they collect personal information after they declare it, well, that’s what happens, whether it’s desirable or not.
It’s give and take. It can’t be helped. They can’t live off of volunteer work either, not even HF.
The question is whether this situation is caused by HF’s conscious default setting, by people or companies who want to collect personal information, or by the fact that HF set the default setting a long time ago and forgot to reconsider it. And furthermore, that they forgot to create an instruction manual or EULA-like document.
I think the latter is more likely.
If HF itself collects personal information for HF, it should be able to track it all without using a gated model, and it should be able to sell it directly to Google or Facebook.