Hacker News Reader

Comment by aucisson_masque | Hacker News Reader

don't know if the Times is such the bad guy that this article presents, imo the justice system in the USA has always been that way. They don't care how extravagant one request is or how little it makes sense so long as you got money and lawyer to push it.

I guess people could switch to one of the many chatgpt competitor that isn't being forced to give away your personal chat.

Don't even know what the time is trying to achieve now, the cat is out of the bag with LLM. Even if a judge ruled that chatgpt and other must give royalties to the time for each request, what about the ones running locally on joe's computer or in countries that don't care about American justice at all.

View on HN

The justice systems hasn't always been quite like this. It's not business as usual for some lawsuit to force a SAAS provider to turn over every scrap of data they stored, even the deleted data, on the off chance it might contain something infringing.

Well, maybe it's business as usual now. A lot of things that were previously considered obvious overreach by corporates and goverment are now depicted as "business as usual" in the US.

They're knowingly contributing to abuse of the discovery process to violate privacy and drive the cost of litigation through the roof. They're absolutely bad guys along with the justice system itself.

> They're knowingly contributing to abuse of the discovery process to violate privacy

Are they? Are you speculating or do you know something we don’t?

It seems that if the NYT want to know whether ChatGPT has been producing copyrighted material, that they own, verbatim, and also understand the scale of the abuse, then they would need to see the logs.

People shouldn’t be surprised, that a company that flouts the law (regardless of what they think of those laws) for its own financial gain, might end up with its collected data in a legal discovery. It’s evidence.

Yeah, the headline is a little bit rage bait. There are countless disclosure requests that happen every day that could be spun to say "X wants your messages from Meta, Twitter etc" Well yeah, this isn't something new.

Infact I see it being hard to defend for OpenAI to basically say "Well yes, its standard practice to hand over any potential evidence, but no we're not doing that".

As for the deleted data, I wonder if legally there are obligations NOT to delete data ?

This isn't a targeted request for one user's Facebook messages. This is a request for every Facebook user's message. And when they don't hand over private DMs, they point and go "especially those, you're obviously hiding something by not giving us the most sensitive user data you have".

It is an absurd breach of user privacy at the scale of tens of millions of Americans that goes well beyond the reasonability for a civil copyright lawsuit.

This sets the precedence that if Gmail gets sued for privacy, now all our emails are leaked. Télécom companies? All of our text messages. Cloud image storage? Woops, gotta hand us every photo any American has taken with a Samsung phone! After all, it might have content that infringes on my copyright!

>Even if a judge ruled that chatgpt and other must give royalties to the time for each request

You don't think that's a victory in and of itself for a business?

Also, you don't need to worry about drug users if you take out the dealer. the users will eventually dry up.

Reply on news.ycombinator.com