The New York Times sues OpenAI, Microsoft over use of its stories to train chatbots

Status
Not open for further replies.
I find that very interesting from a legal standpoint.
Say I take some time and study GRR Martin books and his writing style (he too sued OpenAI in a different lawsuit) or that I indeed study NYT articles and their style. And then I, using the knowledge I obtained by studying them, started producing texts to mimic those styles.

How would that be different from that OpenAI software is doing? From a pure technical standpoint it seems similar.
The only real difference is that I, as one person, can't possibly be a commercial competitor and therefore a commercial threat to NYT. (in theory, I could pose some threat to a lone author like GRRM) But an automated system like OpenAI can study their work much quicker and can reproduce it in much higher numbers. But that in itself is no crime, to be able to do something super quick.

But, in essence, if OpenAI is found guilty, then every human author should be found guilty if they knowingly reproduces someone's style?
But can writing style be copyrighted? Did not GRRM rely on other authors before him to nurture his style? Did not NYT rely on a long history of newspaper standard writing style?

And if the sole sticking point in the case is reproduction where the author knows what he's doing - then how can that be proven beyond doubt? And would then not be enough that OpenAI simply changes its code a bit, so it replies to this kind of a query "Write me an article in GRMM/NYT style" by: "I can not write you in the same style but here is an article in a somewhat similar style. Do decide for yourself if it meets your standards. " But then it provides pretty much the same article as it would have before that disclaimer.

I find the whole thing to be on a very, very shaky legal ground.
 
The New York Times understands copyright law. No one I've seen online does. Copying styles is not the issue. OpenAI has taken military-grade pattern recognition technology and applied it to text and images on a scale far beyond what even a large group of individuals can do, and Microsoft is backing it with billions. Why is that? Because they know they can monetize this across multiple languages around the world. It's about using copyright works without permission or compensation. There's ZERO shakiness about the Times' position, and I'm sure they ran this by their legal department before filing. By the way, I've been working in publishing for decades. Some people want to apply the "we only stole this a little" fallacy. OpenAI took a chance and they are now in the crosshairs, along with Microsoft.
 
edwest4 said:
The New York Times understands copyright law. ... It's about using copyright works without permission or compensation.
Click to expand...
You can copyright a *style*? Can I copyright "rap music" and then sue anyone who produces said music? I'm gonna copyright "science fiction" and sue the bejesus out of anyone producing more crap Star Wars/Trek shows that don;t meet my standards.

In any event, this lawsuit will go nowhere. I've been informed by Real Experts that AI is just a fad with no future; it will surely dry up and blow away before this suit can get anywhere.
 
edwest4 said:
The New York Times understands copyright law. No one I've seen online does. Copying styles is not the issue. OpenAI has taken military-grade pattern recognition technology and applied it to text and images on a scale far beyond what even a large group of individuals can do, and Microsoft is backing it with billions. Why is that? Because they know they can monetize this across multiple languages around the world. It's about using copyright works without permission or compensation. There's ZERO shakiness about the Times' position, and I'm sure they ran this by their legal department before filing. By the way, I've been working in publishing for decades. Some people want to apply the "we only stole this a little" fallacy. OpenAI took a chance and they are now in the crosshairs, along with Microsoft.
Click to expand...
Maybe the media did not convey the issue well, I don't know. If style is not the issue, what is?

You mention "it's about using copyright works without permission/compensation". Okay, so how is OpenAI using NYT works or GRRM works? Is it by reading it? If so, that's legal.
Is it by writing down, word for word, its contect? If so, that's legal.
Is it by memorizing its contents? If so, that too is legal.
Is it doing all that without paying for access? Meaning without subscribing to NYT? That's illegal, of course. Doing it via library would be legal but of course, we know OpenAI doesn't work that way. But solution is pretty simple, to subscribe for access.

So, unless the law changes and explicitly states "Humans reading content is legal. But machine reading content is not legal" then what OpenAI is doing is legal.

Right now, the laws are still possibly behind the times. They may not have been ready for AI changes. But until they change, (and they may change within just years) I am having trouble seeing what law was breached.

In that regard, the NYT lawsuit and other lawsuits may be more about creating awareness and creating pressure on lawmakers to change laws. Rather than actually being about winning this current, specific crop of lawsuits.
 
"I am having trouble seeing what law was breached."

You write a book called Lord of the Rings and apply for a copyright through the Copyright Office. You also add the words "All rights reserved." That means when someone wants to make a movie or TV show based on your book they have to get your permission, and you work out a fair compensation agreement. OR - you can say no. Or say if someone wants to publish a story based on Lord of the Rings. They have to get your permission to publish a story derived from your work - that is called a "derivative work." And you should be paid fair compensation for the publication of that story by another author. OR - you can say no.

In the case of The New York Times, they found articles they had written by using ChatGPT. This COMPUTER PROGRAM required the input of professional writing in order to disassemble it - just like reverse engineering - into its component parts. That way it could characterize "professional writing" as a series of operations derived from the original work - IN ORDER TO MAKE MONEY.

So, to review, the NYT lawsuit is this:

Use of their copyright work without their permission.
Use of their copyright work without fair compensation.
The NYT also seeks the destruction of GPT.
 
edwest4 said:
You write a book called Lord of the Rings and apply for a copyright through the Copyright Office. You also add the words "All rights reserved." That means when someone wants to make a movie or TV show based on your book they have to get your permission, and you work out a fair compensation agreement.
Click to expand...
You write a book about sparkly teen vampires. It's dreadful, but teen and sub-teen girls across the land go nuts for it. A woman decides to write some fan fiction about the main characters doing some BDSM. She decides to publish it... but before going to press, does a bit of a re-write, changing vampires to billionaires, replacing names, etc. But the story remains more or less the same, and of course the origin remains the same. What happens when she publishes? Do you sue her, because she read your stuff and wrote, for profit, something derived from it?

Nope. She goes on to make a ton of money, all legal and aboveboard. Hollywood comes calling and she makes a few more truckloads of cash.

That's what ChatGPT seems to be doing here.
 
Orionblamblam said:
You write a book about sparkly teen vampires. It's dreadful, but teen and sub-teen girls across the land go nuts for it. A woman decides to write some fan fiction about the main characters doing some BDSM. She decides to publish it... but before going to press, does a bit of a re-write, changing vampires to billionaires, replacing names, etc. But the story remains more or less the same, and of course the origin remains the same. What happens when she publishes? Do you sue her, because she read your stuff and wrote, for profit, something derived from it?

Nope. She goes on to make a ton of money, all legal and aboveboard. Hollywood comes calling and she makes a few more truckloads of cash.

That's what ChatGPT seems to be doing here.
Click to expand...
The difference being that Author B paid for Author A's books to read them.

OpenAI/ChatGPT did not pay anyone to use their books/articles.
 
Scott Kenny said:
The difference being that Author B paid for Author A's books to read them.
Click to expand...

You sure? There's these things called "libraries."

And what if Author B picked up a copy of Author A's book on the street? What if she read a pirated copy online? What if she stole it from a bookstore? Even if Author A could prove such a thing, what case would there be to be made?
 
Orionblamblam said:
You write a book about sparkly teen vampires. It's dreadful, but teen and sub-teen girls across the land go nuts for it. A woman decides to write some fan fiction about the main characters doing some BDSM. She decides to publish it... but before going to press, does a bit of a re-write, changing vampires to billionaires, replacing names, etc. But the story remains more or less the same, and of course the origin remains the same. What happens when she publishes? Do you sue her, because she read your stuff and wrote, for profit, something derived from it?

Nope. She goes on to make a ton of money, all legal and aboveboard. Hollywood comes calling and she makes a few more truckloads of cash.

That's what ChatGPT seems to be doing here.
Click to expand...
i auppose that works right up until you call said teen vampire hunter Buffy, Scruffy or Dutch.........

I believe vampire films etc based around the original book etc and using the names/titles had to pay for the privilege.
 
There are those who earn their living by dealing with matters like this. Lawyers. How many of them frequent this forum?
 
Orionblamblam said:
You write a book about sparkly teen vampires. It's dreadful, but teen and sub-teen girls across the land go nuts for it. A woman decides to write some fan fiction about the main characters doing some BDSM. She decides to publish it... but before going to press, does a bit of a re-write, changing vampires to billionaires, replacing names, etc. But the story remains more or less the same, and of course the origin remains the same. What happens when she publishes? Do you sue her, because she read your stuff and wrote, for profit, something derived from it?

Nope. She goes on to make a ton of money, all legal and aboveboard. Hollywood comes calling and she makes a few more truckloads of cash.

That's what ChatGPT seems to be doing here.
Click to expand...
It seems like a good idea, I'm going to start researching about vampire millionaires with BDSM tastes, it will sell well.;)
 
Status
Not open for further replies.

Similar threads

Back
Top Bottom