Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Can I remove my personal data from GenAI training datasets?

knowingmachines.org

Post reply on HN
Print view View original

41–50 of 123 posts

Re: Can I remove my personal data from GenAI training datasets?

#41

Post by edgyquant »

627467 wrote: post #40
esafak wrote: post #33

Earlier quoted context omitted.

The government has no problem getting rid of CSAM and terrorist material, for practical purposes. It's not a technical problem.

It is obviously a technical challenge. Image yourself trying to facing 3 pictures, one of CSAM, another of terrorist material and just some random photo of random guy eating ice scream: what steps would you do to determine if this last photo is "legal" or not? YouTube have decades of attempt at determining if uploaded content violates copyright, normally through fingerprinting content submitted by right holders. YouT…

If you have the image and the person reaches out saying it’s them and they want it removed, you have a pretty good idea of the legality. As stated, this isn’t a technical problem it’s a bureaucratic one. Most companies didn’t think they had to remove it so didn’t have steps in place to do so. Now a lot of them do, with GenAI just being the latest group of companies who feel they have an argument in favor of not having to do so.

Re: Can I remove my personal data from GenAI training datasets?

#42

Post by exabrial »

valianteffort wrote: post #21

I don't get it, you put that information on the internet, you have no expectation to privacy. But now maybe you learned a lesson, and you won't publicly share things you don't want people to see?

> you put that information on the internet, you have no expectation to privacy

That was the expectation until Zuckerberg came along. People were putting tons of stuff on BBSs, intranets, and other stuff 20+ years beforehand because people had the idea of "consent" back then existed: using someones information for a purpose they didn't intend was (and still is) wrong.

Re: Can I remove my personal data from GenAI training datasets?

#43

Post by FeepingCreature »

exabrial wrote: post #42
valianteffort wrote: post #21

I don't get it, you put that information on the internet, you have no expectation to privacy. But now maybe you learned a lesson, and you won't publicly share things you don't want people to see?

> you put that information on the internet, you have no expectation to privacy That was the expectation until Zuckerberg came along. People were putting tons of stuff on BBSs, intranets, and other stuff 20+ years beforehand because people had the idea of "consent" back then existed: using someones information for a purpose they didn't intend was (and still is) wrong.

Using someone's information for a purpose they didn't intend is wrong? What? This just seems obviously mistaken. I don't agree with that at all. I'm not even sure how to argue about that.

Lots of art, science and technology can be considered as "using someone's information for a purpose they didn't intend." It is extremely normal for people to find new uses for things; information is not exempt from this.

Re: Can I remove my personal data from GenAI training datasets?

#44

Post by caturopath »

exabrial wrote: post #42
valianteffort wrote: post #21

I don't get it, you put that information on the internet, you have no expectation to privacy. But now maybe you learned a lesson, and you won't publicly share things you don't want people to see?

> you put that information on the internet, you have no expectation to privacy That was the expectation until Zuckerberg came along. People were putting tons of stuff on BBSs, intranets, and other stuff 20+ years beforehand because people had the idea of "consent" back then existed: using someones information for a purpose they didn't intend was (and still is) wrong.

> That was the expectation until Zuckerberg came along.

This isn't my recollection of a pre-facebook internet

Re: Can I remove my personal data from GenAI training datasets?

#45

Post by vasco »

exabrial wrote: post #42
valianteffort wrote: post #21

I don't get it, you put that information on the internet, you have no expectation to privacy. But now maybe you learned a lesson, and you won't publicly share things you don't want people to see?

> you put that information on the internet, you have no expectation to privacy That was the expectation until Zuckerberg came along. People were putting tons of stuff on BBSs, intranets, and other stuff 20+ years beforehand because people had the idea of "consent" back then existed: using someones information for a purpose they didn't intend was (and still is) wrong.

This seems like a fantasy. Anything could end up in 4chan with a photoshoped penis on it since a long time before Zuck came along.

Re: Can I remove my personal data from GenAI training datasets?

#46

Post by XorNot »

esafak wrote: post #33
qweqwe14 wrote: post #30

Earlier quoted context omitted.

I'm not saying that this right is a bad thing, but the reality is that it's rarely enforceable. As soon as you put something on the internet, you have every "right" to expect that it's there forever. No amount of legislation can outweigh the technical reality.

The government has no problem getting rid of CSAM and terrorist material, for practical purposes. It's not a technical problem.

The government has enormous trouble getting rid of CSAM. The FBI had databases of content that is decades old - basically images recognised as "classics" that are still shared amongst the consumers of such things.

It is illegal to possess, but despite the most aggressive enforcement on the planet it is still out there.

Re: Can I remove my personal data from GenAI training datasets?

#47

Post by Retric »

soulofmischief wrote: post #36
Retric wrote: post #29

Earlier quoted context omitted.

That’s factually incorrect. GitHub’s ToS don’t give them any rights to use your information in this way. Their argument is they don’t need any such permission, but they can profit from this data by selectively allowing large scale scrapping by 3rd parties thus breaking their TOS by selling your data.

GitHub's ToS expressly allows them to change the terms of their service at will, and they do indeed publish these changes, and your continued use of their website is considered consent. Furthermore, as [0] states: > We may use your information to provide, administer, analyze, manage, and operate our Service. For example, we use your information for the following purposes: > ...Improve and develop our products and ser…

They didn’t limit copilot to active users who therefore have implied consent to ToS changes.

“Improving and developing products and services” doesn’t cover what they did. Which is why that line isn’t being brought up in the ongoing litigation.

Further, at scale they are well aware users didn’t own the copyright to all uploaded data making their ToS insufficient shielding for this use. Which is why they aren’t even trying to suggest the ToS enabled this kind of use as that suggests it would be required and opens themselves to massive amounts of liability.

Re: Can I remove my personal data from GenAI training datasets?

#48

Post by Sparkyte »

valianteffort wrote: post #21

I don't get it, you put that information on the internet, you have no expectation to privacy. But now maybe you learned a lesson, and you won't publicly share things you don't want people to see?

GPDR and many data rights exist today, it is still applicable to AI.

Re: Can I remove my personal data from GenAI training datasets?

#49

Post by Sparkyte »

Legally they can't use your data without your consent and that data must be able to be deleted by your demand. So for GenAI to use personal information is bad and they should change its design. I advise they obfuscate the data to not contain sensitive information if they are intending to create AI around such data. Identifiable information is not a good thing.

GDPR and many other laws are still applicable to what GenAI is.

Re: Can I remove my personal data from GenAI training datasets?

#50

Post by happytiger »

graphe wrote: post #38
happytiger wrote: post #17

Earlier quoted context omitted.

Frankly, and not directed personally, but that’s not really true. Government can make the incentive — and has. Legislation can put the teeth in societal goods like this even if financial incentives don’t. California has done exactly this with their right to be deleted law. https://www.foley.com/en/insights/publications/2023/10/calif... .

The two parties are the state and the data brokers. The individual has no way to do this on their own. But you're right, it doesn't have to be true. Maybe someone will have precedence and sue a company. Until then I am convinced it is true.

Fair enough. I respect your logic.
Post reply on HN