Datasets:

alpindale

/

two-million-bluesky-posts

like 191

I personally would much prefer that all of this data be deleted, since there was no request for consent to scrape the posts of any of the people in this dataset. I'm going to assume that many people also share @erisws 's stance on this.

Due to the lack of notification and consent, this dataset may have been obtained illegally according to some consumer data protection and privacy laws, but I haven't yet found comprehensive evidence to confirm or refute this claim.

never mind, it is entirely illegal. see #29 for details.

Bro cannot just make up his own rules and ignore what's not convenient. The screenshot is his opinion, nothing more.

Bro cannot just make up his own rules and ignore what's not convenient. The screenshot is his opinion, nothing more.

I see the only activity on your HuggingFace profile is to comment on this article.

Do you really care to litigate this?

Bro cannot just make up his own rules and ignore what's not convenient. The screenshot is his opinion, nothing more.

I see the only activity on your HuggingFace profile is to comment on this article.

Do you really care to litigate this?

personally I just don't use huggingface much, most of my experimentation on Huggingface involves testing models through the inference API. it's safe to assume they do the same

I will check my options involving taking the data down if my request is ignored.

I see the only activity on your HuggingFace profile is to comment on this article.

Do you really care to litigate this?

What matters to me is that my legal right to control my personal data is respected. I’m not here to discuss whether I’m an active HF user or not.

My formal response to the following:

Bluesky's TOS states quite clearly that content you make you reserve the rights and ownership to, therefore you are still in control of your own data- which is why filing a GDPR request would be viable.

Just because you can freely access the data in an easy-to-use format does not mean that you have the legal right to put it into a public dataset for training models. This issue is similar to scraping images from websites to use them in AI training- however, there is no text equivalent of Nightshade or Glaze to protect what we say.

I uphold my request to have my, and all members of my organization's data removed from this dataset immediately. If ignored, I will pursue other options of having my data taken down.

you guys are being entitled, please realize that anything that can be gathered automatically can be gathered manually, why are you ok with humans looking at your data but you're not ok with a script downloading the data, as long as something is on the public internet, then the cat's out of the bag, be it some guy on huggingface or wayback machine or someone data-hoarding
You're ok with having a few powerful entities like governments or big tech like Google having access to this data and you're not ok with someone making it accessible to everyone (which is the lesser of 2 evils).
I'm with you, it would be nice for me to own anything I put out there on the internet, but I can't, the second best thing is to make sure that everyone has equal access to the internet, not just government web crawlers and big tech companies keeping their gigantic datasets proprietary.

this is part of a much larger discussion about decentralization, but you can't tell people what to do with what they see on the internet, and especially if they didn't AGREE to a TOS, with anything on the internet, the cat's out of the bag. The best we can do is to make sure that everyone has access and not just a select few

you guys are being entitled, please realize that anything that can be gathered automatically can be gathered manually, why are you ok with humans looking at your data but you're not ok with a script downloading the data, as long as something is on the public internet, then the cat's out of the bag, be it some guy on huggingface or wayback machine or someone data-hoarding, you're ok with an organization like google having access to this data and you're not ok with someone making it accessible to everyone (which is the lesser of 2 evils).
I'm with you, it would be nice for me to own anything I put out there on the internet, but I can't, the second best thing is to make sure that everyone has equal access to the internet, not just government web crawlers and big tech companies keeping their gigantic datasets proprietary

I don't want my data gathered manually or automatically for AI training. I post on Bluesky because I want to be able to engage in a polite conversation and post about topics I find nice, not to train another massive AI model to destroy our planet more.

Huggingface already set a precedent against this by deleting their own atempt at this after major community backlash. Someone else doing this should have the same outcome, or at least have people that didn't consent have an option to have their data redacted in a timely manner.

you guys are being entitled, please realize that anything that can be gathered automatically can be gathered manually, why are you ok with humans looking at your data but you're not ok with a script downloading the data, as long as something is on the public internet, then the cat's out of the bag, be it some guy on huggingface or wayback machine or someone data-hoarding, you're ok with an organization like google having access to this data and you're not ok with someone making it accessible to everyone (which is the lesser of 2 evils).
I'm with you, it would be nice for me to own anything I put out there on the internet, but I can't, the second best thing is to make sure that everyone has equal access to the internet, not just government web crawlers and big tech companies keeping their gigantic datasets proprietary

Personal data removal requests must be respected. It's the law that applies to personal data belonging to EU citizens. Google respects it: https://support.google.com/websearch/community-video/280709766/how-to-remove-personal-information-from-google-search-step-by-step-guide

Asking for your legal rights as a private individual to be respected is a basic thing I would really hope everyone could get behind without hours of discussion.

Acting simultaneously condescending, ignorant, and snide about this does not negate the legal impact, nor does it prevent people from filing reports to EU and CA privacy commissions, naming Pygmalion as an interested party. AI developers who act like this are legitimately destroying the reputation of AI and creating a stigma around AI use that will continue to slow its adoption and growth.

This developer is not magically exempt from the bluesky developer terms as well as privacy regulations. You better hope HF deletes this, or it's likely you will face legal issues.

Bro cannot just make up his own rules and ignore what's not convenient. The screenshot is his opinion, nothing more.

I see the only activity on your HuggingFace profile is to comment on this article.

Do you really care to litigate this?

Yes, I do care, because this repository is in direct violation of established privacy laws, and that should not simply be ignored.

I have confirmed that no data on my account is present in this database. I will keep this report up for others to voice their concerns.

Bro cannot just make up his own rules and ignore what's not convenient. The screenshot is his opinion, nothing more.

I see the only activity on your HuggingFace profile is to comment on this article.

Do you really care to litigate this?

Yes, I do care, because this repository is in direct violation of established privacy laws, and that should not simply be ignored.

Go for it!