OpenAI Offers 20 Million User Chats In ChatGPT Lawsuit. NYT Wants 120 Million. (arstechnica.com) 21
An anonymous reader quotes a report from Ars Technica: OpenAI is preparing to raise what could be its final defense to stop The New York Times from digging through a spectacularly broad range of ChatGPT logs to hunt for any copyright-infringing outputs that could become the most damning evidence in the hotly watched case. In a joint letter (PDF) Thursday, both sides requested to hold a confidential settlement conference on August 7. Ars confirmed with the NYT's legal team that the conference is not about settling the case but instead was scheduled to settle one of the most disputed aspects of the case: news plaintiffs searching through millions of ChatGPT logs. That means it's possible that this week, ChatGPT users will have a much clearer understanding of whether their private chats might be accessed in the lawsuit. In the meantime, OpenAI has broken down (PDF) the "highly complex" process required to make deleted chats searchable in order to block the NYT's request for broader access.
Previously, OpenAI had vowed to stop what it deemed was the NYT's attempt to conduct "mass surveillance" of ChatGPT users. But ultimately, OpenAI lost its fight to keep news plaintiffs away from all ChatGPT logs. After that loss, OpenAI appears to have pivoted and is now doing everything in its power to limit the number of logs accessed in the case -- short of settling -- as its customers fretted over serious privacy concerns. For the most vulnerable users, the lawsuit threatened to expose ChatGPT outputs from sensitive chats that OpenAI had previously promised would be deleted. Most recently, OpenAI floated a compromise, asking the court to agree that news organizations didn't need to search all ChatGPT logs. The AI company cited the "only expert" who has so far weighed in on what could be a statistically relevant, appropriate sample size -- computer science researcher Taylor Berg-Kirkpatrick. He suggested that a sample of 20 million logs would be sufficient to determine how frequently ChatGPT users may be using the chatbot to regurgitate articles and circumvent news sites' paywalls. But the NYT and other news organizations rejected the compromise, OpenAI said in a filing (PDF) yesterday. Instead, news plaintiffs have made what OpenAI said was an "extraordinary request that OpenAI produce the individual log files of 120 million ChatGPT consumer conversations."
That's six times more data than Berg-Kirkpatrick recommended, OpenAI argued. Complying with the request threatens to "increase the scope of user privacy concerns" by delaying the outcome of the case "by months," OpenAI argued. If the request is granted, it would likely trouble many users by extending the amount of time that users' deleted chats will be stored and potentially making them vulnerable to a breach or leak. As negotiations potentially end this week, OpenAI's co-defendant, Microsoft, has picked its own fight with the NYT over its internal ChatGPT equivalent tool that could potentially push the NYT to settle the disputes over ChatGPT logs.
Previously, OpenAI had vowed to stop what it deemed was the NYT's attempt to conduct "mass surveillance" of ChatGPT users. But ultimately, OpenAI lost its fight to keep news plaintiffs away from all ChatGPT logs. After that loss, OpenAI appears to have pivoted and is now doing everything in its power to limit the number of logs accessed in the case -- short of settling -- as its customers fretted over serious privacy concerns. For the most vulnerable users, the lawsuit threatened to expose ChatGPT outputs from sensitive chats that OpenAI had previously promised would be deleted. Most recently, OpenAI floated a compromise, asking the court to agree that news organizations didn't need to search all ChatGPT logs. The AI company cited the "only expert" who has so far weighed in on what could be a statistically relevant, appropriate sample size -- computer science researcher Taylor Berg-Kirkpatrick. He suggested that a sample of 20 million logs would be sufficient to determine how frequently ChatGPT users may be using the chatbot to regurgitate articles and circumvent news sites' paywalls. But the NYT and other news organizations rejected the compromise, OpenAI said in a filing (PDF) yesterday. Instead, news plaintiffs have made what OpenAI said was an "extraordinary request that OpenAI produce the individual log files of 120 million ChatGPT consumer conversations."
That's six times more data than Berg-Kirkpatrick recommended, OpenAI argued. Complying with the request threatens to "increase the scope of user privacy concerns" by delaying the outcome of the case "by months," OpenAI argued. If the request is granted, it would likely trouble many users by extending the amount of time that users' deleted chats will be stored and potentially making them vulnerable to a breach or leak. As negotiations potentially end this week, OpenAI's co-defendant, Microsoft, has picked its own fight with the NYT over its internal ChatGPT equivalent tool that could potentially push the NYT to settle the disputes over ChatGPT logs.
This is a bad look for the New York Times (Score:3, Insightful)
NYT's complaint is valid.
That NYT is willing to set the precedent that OpenAI chat logs can be subpoenad is incredibly dangerous. People have all kinds of private conversations with ChatGPT, and this will hurt all of us so that NYT can strike out at LLMs. The trade is not worth it. OpenAI is 100% correct to protect chat data.
Re: This is a bad look for the New York Times (Score:4, Interesting)
Given the nature of OpenAI and their Microsoft masters people should not have been conducting private conversations with ChatGPT to begin with.
I am hardly a fan of NYT but letâ(TM)s be honest here. OpenAI and co. have no qualms about violating user privacy if it so fits them either. Besides their indiscriminate approach to data scraping, Altman and co. have also engaged in questionable ventures such as the WorldCoin scheme that involves scanning your retinas in exchange for digital chuckie cheese tokens. The whole AI industry is built on violating trust and privacy to make money.
So really screw them for trying to wear the flayed skin of digital privacy as a last-ditch defense here.
Re: (Score:3)
NYT's complaint is valid.
That NYT is willing to set the precedent that OpenAI chat logs can be subpoenad is incredibly dangerous. People have all kinds of private conversations with ChatGPT, and this will hurt all of us so that NYT can strike out at LLMs. The trade is not worth it. OpenAI is 100% correct to protect chat data.
But OpenAI has said they delete logs of outputs. If they're now saying that 120 million is too many to go through because more private data could be compromised, that means they haven't been deleting output logs.
Further, since the NYT and others are only looking for articles from their sites, their searches would be limited in nature. This would be no different than you using DDG and searching for how to make a toasted cheese sandwich. You would not get results for how to make a Tokamak reactor (at least
Re: (Score:3)
But OpenAI has said they delete logs of outputs. If they're now saying that 120 million is too many to go through because more private data could be compromised, that means they haven't been deleting output logs.
OpenAI directly lying on something as important as this? Why am I not surprised.
Re: (Score:1)
Re: (Score:2)
Breach of contract? (Score:2)
OpenAI promised to keep those chats private and delete them. Now after it said deleted them, they're saying they can make them searchable?
If a court is forcing anything here, it should be that those chats stay "deleted".
It is a good lesson that whatever you send off to a cloud service can forever be retained, despite agreements to the contrary.
Re: (Score:3)
The judge is an 80 year old who was appointed by Democrat Bill Clinton in 1995.
People of his demographic cherish the New York Times and don't understand the first thing about AI, cloud computing or really any kind of computing.
OpenAI is going to have a tough run at this. The judge's innate biases will certainly play a role in how he rules on the motions of the case.
Re: Breach of contract? (Score:2)
Seems like this case highlighted a dodgy practice from OpenAI, where they said a conversation would be deleted, but decided to keep a copy and just tell the user it was deleted.
OpenAI should be forced to permanently delete the chats. They probably kept them for future training. Now their breach of privacy is coming out
Re: (Score:2)
Re: (Score:2)
The part where they were told by OpenAI that the logs would be deleted.
You know how when someone says something and you both agree, and your actions going forward are based on those agreements? That thing some people call a contract.
Re: (Score:2)
Even 80 y/o judges know a fishing expedition (Score:2)
They don't need to know those things. My father was retired federal law enforcement and that age. If he were around to read this, he'd lose his shit on that judge for authorizing the most insane fishing expedition ever authorized by an idiotic federal judge.
The judge isn't ignorant of tech. He's straight up stupid, incompetent or corrupt. He should be impeached
Re: (Score:2)
...for this level of not even first year law school tier bullshit.
When someone gets sued (possibly even prior to the suit being filed) in the U.S., he is under an automatic legal obligation to retain ALL records for Discovery. Failure to do so has some serious legal repercussions. So no, this judge is not doing anything out of the ordinary. He is simply requiring OpenAI to abide by its standard legal obligations.
Re: (Score:2)
Re: (Score:2)
No. From an article on the 5th of June:
OpenAI is now fighting a court order to preserve all ChatGPT user logs—including deleted chats and sensitive chats logged through its API business offering
The 20 million record offer describes they need to retrieve chat logs from their offline storage, containing "tens of billions of logs". It appears "deleted" only means online copies are deleted.
NY Times (Score:2)
Of all places The NY Times is going to get a look at deleted logs. If any NY Times employee uses the information in that, they should be sent to jail for a while.