Will the NYT lawsuit against OpenAI and MSFT go to trial?
187
550
1.7k
2025
50%
chance

NYT has sued OpenAI for using its articles for training ChatGPT: https://www.bbc.co.uk/news/technology-67826601

The article states that NYT has already tried to resolve the matter out of court since April.

The steps in a civil lawsuit are:

  1. Pleadings (Initial complaint filed by plaintiff (done), defendant answers or files a motion to dismiss)

  2. Discovery

  3. Trial

  4. Verdict

This question resolves no if a settlement is reached out of court before opening statements are made in a trial (step 3), or the case is dismissed pre trial. Edit: any sequence of events that avoids any trial resolves no.

The question resolves yes if opening statements are presented in a trial, even if a settlement is reached later. Edit: a trial for a subset of the complaints or to determine damages also resolves as yes.

If neither settlement nor dismissal occurs and trial has not started by the closing date (end of 2024) the question resolves n/a

Get Ṁ600 play money
Sort by:

it’s way too transformative for them to have any legitimate copyright claim. If krugman reads the WSJ and incorporates that info into his shitty columns, WSJ doesnt’t get additional revenue proportional to the number of krugman readers. ChatGPT is doing the same thing but better.

predicts NO

@JonathanRay That’s an interesting way to look at it. They do seem to be claiming that you can get ChatGPT to spit out parts of subscriber only articles, which I couldn’t do with straightforward questions.

Generally speaking I’d expect a lot of legal posturing and filing lawsuits to force settlements that would be cheaper than a full trial, but on the other hand this one has such huge implications to every LLM that I can’t imagine settling is a realistic option for OAI —it would open up the floodgates.

So I dunno…

Well, hey, OpenAI clearly stole the NYT content model of hallucinating things and asserting they were true. It was settle, sue, or hire GPT4 as a senior editor to succeed Andrew Rosenthal.

bought Ṁ10 of NO

@SEE That was cool to learn about. Thanks for linking!

2024 is too soon a closing date. Complex civil cases typically take several years before resolution, if they go to trial.

Also, please discuss the market resolutions under the following circumstances:

  • Summary judgment for defendants (not dismissal)

  • Summary judgment for plaintiffs as to liability and damages

  • Summary judgment as to liability; trial as to damages

  • Stipulated set of issues for trial

  • Dismissal that is reversed on appeal

  • etc.

predicts NO

@octothorpe I need to look into the chances that we know whether the case is going to trial or not within 12 months. If there isn’t a realistic chance we do I can move the resolution date if nobody objects. Otherwise, if there is a significant chance of trial beginning within that timeframe I can keep it as is and resolve n/a in the small chance we don’t know.

All summary judgements that avoid a trial resolve this as No.

Partial summary judgment that leave issues to be tried do not resolve this. And a trial to adjudicate a subset of the complaints or to determine damages resolve this as yes. I will note that opening statements in a trial have to occur to resolve this as yes. Any outcome that avoids any trial, even at the last minute, resolves this as no. Makes sense?

predicts NO

@octothorpe Note that the question resolves yes as soon as a trial starts, and doesn’t depend on the trial concluding in any way.

More related questions