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

Fighting the New York Times' invasion of user privacy

openai.com

Post reply on HN
Print view View original

411–420 of 441 posts

Re: Fighting the New York Times' invasion of user privacy

#411

Post by johnwheeler »

code_for_monkey wrote: post #395
johnwheeler wrote: post #141

Earlier quoted context omitted.

Exactly. And the OpenAI corporates speak acting like they give a shit about our best interests. Give me a break, Sam Altman. How stupid do you think everyone is? They have proven that they are the most untrustworthy company on the planet And this isn't AI fear speaking. This is me raging at Sam Altman for spreading so much fear, uncertainty, and doubt just to get investments. The rest of us have to suffer for the las…

You got downvoted for this? That many people are doing 'Leave Sam Altman alone!'? kinda wild

It's weird. It went up and down and up and down. Controversial POV. But thanks for the support. Sam Altman's just too dishonest. It's been said time and time again by so many people, by Paul Graham, Ilya Sutskeve, everybody's telling everybody he's dishonest. When are we going to wake up and get this guy out of there?

Re: Fighting the New York Times' invasion of user privacy

#412

Post by freejazz »

keeda wrote: post #410
freejazz wrote: post #403

Earlier quoted context omitted.

Primarily you seem to be missing the fact that the NYT case is about outputs, not just the training.

Hmm, this is an interesting framing of the lawsuit. If it's about outputs and not just training, are the outputs really orthogonal to the training? In traditional computer systems, no, outputs are always a function of inputs. LLMs throw a wrench into this reasoning because they apply opaque statistics to a combination of training data and the user prompt to produce outputs, so the input-output relationship is much le…

>Hmm, this is an interesting framing of the lawsuit.

First, it's not a "framing" of the lawsuit. A lawsuit is a number of claims made by one party against the other. In the two California cases, there were no decisions made on claims relating to LLM outputs. In the NYT case, there are claims relating to LLM outputs.

Yes, it could also be about training. But the discovery pertains to the outputs, which is the issue in this case. So even if you apply the holding that training is fair, which I don't see likely to happen in the district courts of the second circuit, you still don't get the result that the person I responded to suggested, which was that this should all be moot because of two decisions in two different cases in California which are not binding precedent in the 2nd circuit, and which also would not dispose of all of NYT's claims.

>So then this case should also be about training. The question then is: did OpenAI intend to have these models be able to regurgitate large amounts of content? Or is it yet another emergent property that nobody anticipated?

Intent is not a required element of copyright infringement, so you'd be wrong there. Plaintiffs can use intent to evidence willful infringement, which they are entitled to do in statutory damages cases, and receive a damages multiplier, which this one is. So OpenAI can't avoid liability based on their intent or a lack thereof. They can only, at best, use 'intent' to establish that NYT's is not entitled to heightened damages.

>So this might come down to intent.

It's always amusing to see people apply completely made up rationales to legal cases based upon their own personal feelings about technologies while completely disregarding, lets say, 100 years of legal jurisprudence.

Re: Fighting the New York Times' invasion of user privacy

#413

Post by rpdillon »

otterley wrote: post #380
rpdillon wrote: post #336

Earlier quoted context omitted.

This response is misleading. Almost all computer services keep logs for a short period of time, so the court order to retain existing information is quite a bit more powerful than a layman would think. Because a huge amount of data is retained for a short period of time and then rapidly deleted in most web services I've worked on for the past 30 years. This is true in services like Datadog, New Relic, and logging ser…

There is an important distinction that relates to a court’s ability to order a defendant to perform work to facilitate discovery. A court can order preservation of records, but they generally cannot order a defendant to create new ones. I was responding to your use of the word “collect,” which implies significantly more effort than merely not destroying logs (i.e. logging new information that they weren’t already). I…

Both are simply software changes. In one case, they're going to have to alter the software to not delete chats that users request to be deleted. In the other case, they'll alter the software to log new information. Neither of these are particularly difficult.

Re: Fighting the New York Times' invasion of user privacy

#414

Post by stevage »

rpdillon wrote: post #357
stevage wrote: post #287

As soon as I see someone claiming a lawsuit against them is "baseless" I'm deeply sceptical about everything that follows.

When I looked for the base of this lawsuit, I was looking for some kind of monetary damage that the New York Times had suffered as a result of open AI's actions, like specific cases where their work has been reproduced or people canceling their subscriptions to the New York Times because of OpenAI's launch. I've done so much reading, and I've still been unable to find anything that articulates this. Do you know of an…

>specific cases where their work has been reproduced

Isn't that exactly what they're trying to find by looking through OpenAI customers' conversations?

Re: Fighting the New York Times' invasion of user privacy

#415

Post by otterley »

rpdillon wrote: post #413
otterley wrote: post #380

Earlier quoted context omitted.

There is an important distinction that relates to a court’s ability to order a defendant to perform work to facilitate discovery. A court can order preservation of records, but they generally cannot order a defendant to create new ones. I was responding to your use of the word “collect,” which implies significantly more effort than merely not destroying logs (i.e. logging new information that they weren’t already). I…

Both are simply software changes. In one case, they're going to have to alter the software to not delete chats that users request to be deleted. In the other case, they'll alter the software to log new information. Neither of these are particularly difficult.

I understand, but the law still distinguishes between the two cases. In my experience, typically expunging is handled by a process separate from its creation (it depends on the logging framework, of course). And with the increasing trend of generated logs being ingested, processed, and stored by separate services, often disabling log deletion is a mere API call away.

Re: Fighting the New York Times' invasion of user privacy

#416

Post by fenomas »

gausswho wrote: post #384
fenomas wrote: post #304

Earlier quoted context omitted.

Chatgpt.com is essentially a CRUD app. What you're saying here amounts to saying that it could conceivably have been designed to work dramatically differently from all other CRUD apps. And obviously that's true, but why would it be? It's a website! You submit text, that you'll view or edit later, so the server stores it. How is that controversial to a HN audience? Also: > the clients don't need to be running at the s…

> An always-on device that stores data in order to sync it to clients is a server. Yes. But it's my server. I burden myself to operate it so that persistence does not come at the cost of control. I think we might be tilting at different windmills here.

TBH it sounds like you're just imagining a very different service than the one openAI operates. You're imagining something where you send an input, the server returns an output - and after that they're out of the equation, and storing the output somewhere is a separate concern that could be left up to the user.

But the service they actually operate is functionally a collaborative document editor - the chat histories are basically rich text docs that you can view, edit, archive, share with others, and which are integrated with various server-side tools. And the document very obviously needs to be stored on the server to do all those things.

Re: Fighting the New York Times' invasion of user privacy

#417

Post by keeda »

freejazz wrote: post #412
keeda wrote: post #410

Earlier quoted context omitted.

Hmm, this is an interesting framing of the lawsuit. If it's about outputs and not just training, are the outputs really orthogonal to the training? In traditional computer systems, no, outputs are always a function of inputs. LLMs throw a wrench into this reasoning because they apply opaque statistics to a combination of training data and the user prompt to produce outputs, so the input-output relationship is much le…

>Hmm, this is an interesting framing of the lawsuit. First, it's not a "framing" of the lawsuit. A lawsuit is a number of claims made by one party against the other. In the two California cases, there were no decisions made on claims relating to LLM outputs. In the NYT case, there are claims relating to LLM outputs. Yes, it could also be about training. But the discovery pertains to the outputs, which is the issue in…

Oh I'm totally an armchair lawyer, so my ruminations were not grounded in laws or legal precedence :-) I do have some background on the patent side of things, where independent reinvention is also not a defence for infringement, but not so much in copyright, so this was educational.

However, has there been any case where the infringment was not only unintentional, but also unexpected?

That is, if you look at cases of uintentional infringement, these are typically cases where some the act of reproduction of content was intentional, but there was a lack of awareness or confusion about the copyright protections of that content. (This paper was useful for background: https://www.law.uci.edu/faculty/full-time/reese/reese_innoce...)

But I could not find a case where the act of copying itself was non-intentional.

In this case, looking at how LLM training works and what LLMs do, it is surprising that it could reproduce the training content verbatim. The fact that it reproduced those outputs is undeniable, but how does existing law and jurisprudence apply to an unprecedented case like this where the reproduction was through some magic black box that nobody can decipher?

Re: Fighting the New York Times' invasion of user privacy

#418

Post by frm88 »

rpdillon wrote: post #356
frm88 wrote: post #314

Earlier quoted context omitted.

to utilize intellectual property from others (either directly in the model aka NYT, or indirectly via web searches) without any rights ... and put the liability for retrieving said property and hence the culpability for copyright infringement on the enduser: Since the output would only be generated as a result of user inputs known as prompts, it was not the defendants, but the respective user who would be liable for…

But wait, isn't this what we want? This means the models can be very powerful and that people have to use their judgment when they produce output so that they are held accountable for whether or not they produced something that was infringing. Why is that a bad thing?

Can I ask you why we would the enduser be punishable for the pirating OpenAI did? That would mean governments have to take the next step to protect copyrighted material and what we face then I don't even dare to imagine.

Re: Fighting the New York Times' invasion of user privacy

#419

Post by cyrusradfar »

The request is for redacted logs. Redaction must be reasonable but OpenAI can protect the PII across 20M conversations if they:

  1. search conversations for PII they know of for the given user and redact

  2. use their own models to scrub the conversations of non-verbatim mentions
In the end, the NYT isn't asking for the identity of the posters. If that was the case, I'd 100% be onboard to scream bloody murder with 'em.

Re: Fighting the New York Times' invasion of user privacy

#420

Post by Aerroon »

freejazz wrote: post #375
Aerroon wrote: post #326

Earlier quoted context omitted.

> The only potential issue is them finding something juicy in someone's chat, that they could publish as a story; and then claiming they found out about this juicy story through other means, (such as a confidential informant) Which is concerning since this is a news organization that's getting the data. Let's say they do find some juicy detail and use it, then what? Nothing. It's not like you can ever fix a privacy v…

>Let's say they do find some juicy detail and use it, then what? Nothing. It's not like you can ever fix a privacy violation. Nobody involved would get a serious punishment, like prison time, either. There are no privacy violations. OpenAI already told the court they anonymized it. What they say in court and what they say in the blog is different and so many people here are (unfortunately) falling for it!

There's no such thing. Anonymized data can still be used to identify someone as we've seen on numerous occasions.
Post reply on HN