Ethical question as it pertains to AI documents on Word

[WaikikiFirst]

let's look at the definition of plagiarize via post #12.

I can't say I agree with that definition, esp since it referred to copyright and you refer to plagiarize. So much confusion.

Screen scraping, which is the term you're referring to, really isn't applicable specific to LLMs

How much do you want to bet that one of the main arguments of the plaintiffs will be that the typical "AI Overview" output is intellectually a minor step above screen-scraping? Do you think the journalists who used the term "scrape" are CS majors? They most likely have seen the lawsuit as it now stands and got the word from that.
"Screen"-scraping is not the only type of "scraping" done in the computer world. I made a ton of money once shorting a software stock that really just did 2 other kinds of scraping. The people (like Jim Cramer) who didn't get that thought it would be a key player in 2 huge markets and bid the stock way up. It didn't take 6 months for Cramer to be proven dead wrong, and eventually it got sold for the patents. I would say "scraping" is basically just re-formatting and, yes, an LLM does more than that. The plaintiff's lawyers will argue it does LITTLE more than that.

 

[ScoopKona]

I agree with the notion that we're already too far gone for this to be reined in.

AI is going to calcify society even more. There will be the people who don't need AI at the top. And the people who can't do anything at all without AI at the bottom. The people who need to be spoon-fed will become a chronic consumer underclass. Sheep to be sheared. We're already seeing this.

 

[jp10558]

Such people do not have a commercial relationship with the producer of the documents, other than maybe having bought 1 or 2 copies.

Isn't a major thrust of the lawsuit that OpenAI etc don't have a relationship (and the producers wish they did a la buying some special license) and instead just read everything publicly available online?

Copyright, a commercial concept, is a better path for discussion than is "plagiarism", which is more of an academic concept. I wonder if all the lawyers who have specialized in "Digital Law" for the last 30 years think it is crazy.
Never underestimate how the digital world differs from the physical world, nor whether lawyers will get involved.

If any of the law made any internally consistent or logically consistent sense, we'd be in a better place. But it doesn't. So I have no idea how it'll apply, I'm just making a point that you can find cases going all sorts of ways and we won't know, and I doubt can meaningfully predict what would happen.

 

[jp10558]

You KNOW this? Really? I'm shocked. I'd love to see your research.

Well, KNOW is a strong term. I think it's likely that it wouldn't be news the few times someone manages to get an AI to reproduce copyrighted material if it wasn't pretty rare. I certainly don't see that happen in my use. Most people I've seen report about it don't see it either.

 

[jp10558]

I would say "scraping" is basically just re-formatting and, yes, an LLM does more than that. The plaintiff's lawyers will argue it does LITTLE more than that.

Sure, but the defendants can certainly argue a number of things. Like, many searches are for basic facts. You can't copyright the phone numbers to people in the phone book for instance, and I'd be surprised if stuff like the hours of operation or location could be restricted either. The only thing there you COULD copyright is the one thing that's gotten rid of, the formatting. While I'm less sure - I would expect to see things like the basic public description of the business as provided by the business also can't be copyrighted really either - at least the basic underlying facts like "Starbucks sells coffee and is a chain" sort of thing.

On the other side of course are things like https://www.cnn.com/2015/05/27/living/richard-prince-instagram-feat where someone straight up stole others photos and sold them (for LOTS of money) and was found not to be breaking copyright. So clearly re-presenting pictures for instance isn't inherently a copyright violation.

I bet we'll get some decision in 10 years that'll be irrelevant as the world will have moved on.

 

[WaikikiFirst]

Isn't a major thrust of the lawsuit that OpenAI

gawd, you got me. I haven't researched a lawsuit that was just filed yesterday, but wth makes you think OpenAI is any part of that lawsuit? more generally, I think you don't understand the commercial relationships in the search / digital-ads market AT ALL. I know I have followed it for 20 yrs and I don't really understand them, since I have never seen the contracts nor worked at one of the "exchanges", but you ... you just keep going on about things you seem to know nothing about

I doubt can meaningfully predict what would happen.

I can agree with that. "The Markets" are constantly groping to figure out wth these judges are doing with both their initial decisions and then their "remedies".

 

[WaikikiFirst]

Well, KNOW is a strong term. I think it's likely that it wouldn't be news the few times someone manages to get an AI to reproduce copyrighted material if it wasn't pretty rare. I certainly don't see that happen in my use. Most people I've seen report about it don't see it either.

astounding. Water isn't wet. The sky isn't blue. Snow isn't white. Frank Zappa agrees with you there

 

[jp10558]

gawd, you got me. I haven't researched a lawsuit that was just filed yesterday, but wth makes you think OpenAI is any part of that lawsuit?

Sorry, I meant the ongoing NYT etc vs OpenAI lawsuit that is about copyright. I have no idea about a lawsuit filed yesterday and was not intending to reference that.

more generally, I think you don't understand the commercial relationships in the search / digital-ads market AT ALL. I know I have followed it for 20 yrs aI can agree with thnd I don't really understand them, since I have never seen the contracts nor worked at one of the "exchanges", but you ... you just keep going on about things you seem to know nothing about

Maybe you're just very unclear about what the heck you're talking about. I have been referring to copyright and ethical / philosophical thoughts. No contract, no commercial relationships etc. So yea, if you apply a random completely different lens, I'll look insane in what I'm saying. As far as I know, at best I understood "commercial relationships" to be about licensing in the context of copyright. I guess sure - they could also sue for breach of contract or something but I wasn't thinking about that from "ethical concerns". We've already agreed (I think) that systems of ethics have little relationship to the law, and probably even less to random contracts different people may have.

 

[jp10558]

astounding. Water isn't wet. The sky isn't blue. Snow isn't white. Frank Zappa agrees with you there

Well, let's just say it was a lot more obvious to everyone that Napster reproduced copyrighted material back in the day.

 

[WaikikiFirst]

I have been referring to copyright and ethical / philosophical thoughts.

Yes, you are not the only one here blurring the otherwise clear line between ethics and the commercial, economic bases of copyright law. I'm sure they teach that all "law" has an "ethical" basis, but you're just refusing to see the difference in real-world application.

 

[jp10558]

Yes, you are not the only one here blurring the otherwise clear line between ethics and the commercial, economic bases of copyright law. I'm sure they teach that all "law" has an "ethical" basis, but you're just refusing to see the difference in real-world application.

And I feel like you've completely missed the first post and ostensible topic of the thread. No where did the discussion starter talk about legal at all. So to expect all posts to be about strictly legal ideas is pretty wrong IMO. I think we violently agree that legally it's a huge grey muddy area that no one has any idea how courts will rule. So copyright out of the way, let's discuss the live issue. Ethical positions can be more interesting to discuss - how things "should" be. And my position is that I don't personally see why changing from a human to a computer should change what's OK to do.