3 ms·

If you wrote a book on a very rare subject that you were basically the only popular source of detailed info on... And now these models just answer questions for

If you wrote a book on a very rare subject that you were basically the only popular source of detailed info on... And now these models just answer questions for people without having to buy the book.. is that really fair? What if a person bought a copy of the book, read it and learned the material, and then provided advice to people based on the material. I think most people would say that that is fair. So why is it less fair for the AI to do what is effectively the same thing? Is it simply a matter of scale? Or is there something more fundamental at work?

nicenewtemp84 4y ago
That was basically my edit, we may have missed each other. The only point I can see making is that for some reason I doubt openai actually paid the author for the contents of the book, but maybe they did. Maybe they borrowed ebooks one by one from some digital library and ingested them that way for free.
- arthurcolle 4y ago
  Somehow I feel like they just scraped libgen and sci hub Much more "scalable" i.e. easy to get raw training dataset. Move fast and break things
mywittyname 4y ago
I'd argue ChatGPT is closer to restating copyrighted materials in a different way. And that is legally questionable. Whether that actually matters depends on how much money is at stake and how much each party cares to spend to enforce or protect their viewpoint. Allowing these systems to utilize copyrighted materials without compensating the copyright holder is a legal loophole the size of the sun. Then it becomes a perfectly legal defense to write "an AI" that slurps up copyrighted material and produces results that are "based on" that material. My issue is that the copyrighted work is fundamentally integrated into the algorithm. Say you wrote a program that would input a copyrighted work, then prints out the contents, but with some words replaced with synonyms. Would the output of that be okay? What about if phrases were replaced? How about if entire sections are added, rearranged, and/or removed? At some point, the results become so mangled as to not be recognizable, but the fact that a copyrighted work serves as an input to the program is still an issue, IMHO. A major difference between humans and computers is that a human can't (generally) recall hundreds of pages of text and regurgitate it on demand. So a human who has read a book only retains a fraction of the contents, and even then, generally they don't retain all of the information accurately.
- mindcrime 4y ago
  I get what you're saying, but I'm still not convinced that introducing the notion of "a computer program" vs "a human" changes anything in the equation. That is to say, IF you - as a human - had a "photographic memory" and after reading a copyrighted work began to spit out big chunks of it verbatim in response to a query, then you would already be guilty of copyright infringement, no? Likewise, if a computer program reads 10 books on, I dunno, let's say Quantum Physics, and then - in response to a query - emits a technically correct answer that isn't obviously cribbed from any one book... has it violated copyright? My point is not to say that contemporary "AI" systems never emit things that are arguably copyright infringement. It's more just to say that there's nothing about being an AI that makes it necessary that its output be copyright infringing.
- amanaplanacanal 3y ago
  As far as I can tell, restating materials in a different way does not infringe copyright.
  - SilasX 3y ago
    Right, that’s just citing sources. Even if your source is only one book (the Cliff Notes model).
  - Engineering-MD 3y ago
    It can be seen as a derived work. A translation of Harry Potter in to French, Japanese or Latin is not free from the copyright of the original work. That is literally restating materials in a different way.
tonyedgecombe 4y ago
I do wonder if it will disincentivise the author from writing the book in the first place. What’s the point if an AI is going to regurgitate the material to every Tom, Dick and Harry thereby killing the demand for the book.
exabrial 4y ago
Answer is pretty obvious: One is a computer program, literally designed to pilfer copyrighted works, the other is a human being.
- function_seven 4y ago
  If you smuggle "pilfer" into the description of AI, then in that case it's "obvious". But the comparison could just as easily be written, "One is a computer program, literally designed to learn from copyrighted works, the other is a human being that learns from copyrighted works" If ChatGPT is parroting entire paragraphs of copyrighted work in its answers, then "pilfer" is probably the right word. Or maybe "plagiarize". But if it's training on the information gleaned from the work, and using that training to synthesize the answers, isn't that a lot more like how a learned person applies their knowledge? I haven't yet made up my mind completely on how IP should interact with AI. I see good arguments for and against. But the answer is far from obvious!
- mindcrime 4y ago
  Why is it "pilfering" when the computer program does it, but "learning" when a human does it?
  - exabrial 4y ago
    Because those two things are not even equivalent. A human can reason, make decisions, act in best interest, judge context of a situation, apply it to new scenarios, judge emotion, and create a new derivative work. The computer program is literally taking someones work and adding fancy way to search through the other person's work, without adding any value. We need to stop fantasizing these fancy toys are actually any sort of intelligence.
    - mindcrime 3y ago
      We need to stop fantasizing these fancy toys are actually any sort of intelligence. That strikes me as a rather extraordinary claim. It seems obvious to me that contemporary AI's are "some" sort of intelligence, albeit with open questions around "what kind of intelligence?" and "how intelligent are they?".
JonChesterfield 3y ago
What if one person bought the book, copied it, and distributed it for negligible cost to anyone with a passing interest in the contents? Sounds pretty similar to copyright infringement.

nicenewtemp84 4y ago

That was basically my edit, we may have missed each other. The only point I can see making is that for some reason I doubt openai actually paid the author for the contents of the book, but maybe they did. Maybe they borrowed ebooks one by one from some digital library and ingested them that way for free.

arthurcolle 4y ago
Somehow I feel like they just scraped libgen and sci hub Much more "scalable" i.e. easy to get raw training dataset. Move fast and break things

mywittyname 4y ago

I'd argue ChatGPT is closer to restating copyrighted materials in a different way. And that is legally questionable. Whether that actually matters depends on how much money is at stake and how much each party cares to spend to enforce or protect their viewpoint. Allowing these systems to utilize copyrighted materials without compensating the copyright holder is a legal loophole the size of the sun. Then it becomes a perfectly legal defense to write "an AI" that slurps up copyrighted material and produces results that are "based on" that material. My issue is that the copyrighted work is fundamentally integrated into the algorithm. Say you wrote a program that would input a copyrighted work, then prints out the contents, but with some words replaced with synonyms. Would the output of that be okay? What about if phrases were replaced? How about if entire sections are added, rearranged, and/or removed? At some point, the results become so mangled as to not be recognizable, but the fact that a copyrighted work serves as an input to the program is still an issue, IMHO. A major difference between humans and computers is that a human can't (generally) recall hundreds of pages of text and regurgitate it on demand. So a human who has read a book only retains a fraction of the contents, and even then, generally they don't retain all of the information accurately.

mindcrime 4y ago
I get what you're saying, but I'm still not convinced that introducing the notion of "a computer program" vs "a human" changes anything in the equation. That is to say, IF you - as a human - had a "photographic memory" and after reading a copyrighted work began to spit out big chunks of it verbatim in response to a query, then you would already be guilty of copyright infringement, no? Likewise, if a computer program reads 10 books on, I dunno, let's say Quantum Physics, and then - in response to a query - emits a technically correct answer that isn't obviously cribbed from any one book... has it violated copyright? My point is not to say that contemporary "AI" systems never emit things that are arguably copyright infringement. It's more just to say that there's nothing about being an AI that makes it necessary that its output be copyright infringing.
amanaplanacanal 3y ago
As far as I can tell, restating materials in a different way does not infringe copyright.
- SilasX 3y ago
  Right, that’s just citing sources. Even if your source is only one book (the Cliff Notes model).
- Engineering-MD 3y ago
  It can be seen as a derived work. A translation of Harry Potter in to French, Japanese or Latin is not free from the copyright of the original work. That is literally restating materials in a different way.

tonyedgecombe 4y ago

I do wonder if it will disincentivise the author from writing the book in the first place. What’s the point if an AI is going to regurgitate the material to every Tom, Dick and Harry thereby killing the demand for the book.

exabrial 4y ago

Answer is pretty obvious: One is a computer program, literally designed to pilfer copyrighted works, the other is a human being.

function_seven 4y ago
If you smuggle "pilfer" into the description of AI, then in that case it's "obvious". But the comparison could just as easily be written, "One is a computer program, literally designed to learn from copyrighted works, the other is a human being that learns from copyrighted works" If ChatGPT is parroting entire paragraphs of copyrighted work in its answers, then "pilfer" is probably the right word. Or maybe "plagiarize". But if it's training on the information gleaned from the work, and using that training to synthesize the answers, isn't that a lot more like how a learned person applies their knowledge? I haven't yet made up my mind completely on how IP should interact with AI. I see good arguments for and against. But the answer is far from obvious!
mindcrime 4y ago
Why is it "pilfering" when the computer program does it, but "learning" when a human does it?
- exabrial 4y ago
  Because those two things are not even equivalent. A human can reason, make decisions, act in best interest, judge context of a situation, apply it to new scenarios, judge emotion, and create a new derivative work. The computer program is literally taking someones work and adding fancy way to search through the other person's work, without adding any value. We need to stop fantasizing these fancy toys are actually any sort of intelligence.
  - mindcrime 3y ago
    We need to stop fantasizing these fancy toys are actually any sort of intelligence. That strikes me as a rather extraordinary claim. It seems obvious to me that contemporary AI's are "some" sort of intelligence, albeit with open questions around "what kind of intelligence?" and "how intelligent are they?".

JonChesterfield 3y ago

What if one person bought the book, copied it, and distributed it for negligible cost to anyone with a passing interest in the contents? Sounds pretty similar to copyright infringement.