5 ms·

This is an incredibly biased article, hinging entirely on the assumption that AI training is fair use.

by jprete 3y ago

zinekeller 3y ago
Unfortunately it seems majority of HN clings on "AI is fair use!" the same way GIFs (of movie and TV snippets) are "perceived to be fair use" (but is simply not fair use in most cases).
- 542458 3y ago
  Why do you feel that the Google vs Authors Guild ruling cited in the article isn’t relevant?
- robertlagrant 3y ago
  "majority" and "clings on" seem entirely arbitrary and perjorative. Why not make your actual point instead of just beating up a straw man?
- luma 3y ago
  And the other half of HN loses their mind when AI is brought up, completely missing the point made quite clearly in the OP that _no_AI_was_involved_in_any _of_this_. The AI hysteria on this site is ridiculous.
  - quickthrower2 3y ago
    The ai moving goal! No 2023.ai was used. Maybe some 1993.ai. Chat GPT will be considered “just stats” in 2033.
pmarreck 3y ago
There is no “AI” in the project. It literally just seems to do Bayesian sentiment analysis of books. It does NOT then mine that data to score OTHER books. Please actually read the article next time before commenting
- jprete 3y ago
  I did in fact read the article. The author isn’t concerned about the book project. The author wants to use it as a springboard for complaining about AI fear, and his argument fundamentally hinges on the claim that AI training is fair use. He’s claiming the book project is comparable to AI training. I think the book project has a better claim to fair use than AI training does, although excerpting entire pages is maybe too far (but the word highlighting does feel like commentary).
  - getabrainplz 3y ago
    If I read the book, take notes about the number of times $x happened and later post them to my website, are you fine with that? if I automate that process with a script that uses regular expressions, are you fine with that? if I add a neural network to my script and it spits out the same numbers, are you still fine with that? hopefully you see the madness in your argument.
    - jprete 3y ago
      I’m thinking specifically of generative AI’s fair use claims. I thought the book site (and all of your extension examples) was clearly fair use.
542458 3y ago
The article cites the Google vs Authors Guild case (https://www.techdirt.com/2013/11/14/google-gets-total-victory-over-authors-guild-book-scanning-is-fair-use/ https://www.techdirt.com/2013/11/14/google-gets-total-victor...) which was a total victory for Google. This seems fairly conclusive that the textual analysis here is fair use to me. > Similarly, Google Books is also transformative in the sense that it has transformed book text into data for purposes of substantive research, including data mining and text mining in new areas, thereby opening up new fields of research. Words in books are being used in a way they have not been used before. Google Books has created something new in the use of book text — the frequency of words and trends in their usage provide substantive information. Furthermore, is this actually AI training? This just looks like stats based on heuristics to me, I.e., garden variety sentiment analysis.
- mycologos 3y ago
  I think Google Books is cool, but "the frequency of words and trends in their usage provid[ing] substantive information" pre-dates Google Books by a long time. For example, there's a collection of word frequencies in the complete works of John Keats from 1917 [1]. Manually tabulated, too! [1] https://catalog.hathitrust.org/Record/001023999 https://catalog.hathitrust.org/Record/001023999
  - 542458 3y ago
    Sure, but the Google Books case is a massive and well-funded court case that was a pretty resounding victory for Google and textual analysis of copyrighted works in general, so anybody arguing that this is obviously copyright violation needs to explain why the Google Books case isn’t relevant.
js8 3y ago
Why it wouldn't be? Indexing is fair use. I think where it gets murky with fair use would be if AI could actually plagiarize the book, but other than that, it should be fair use.
- Nullabillity 3y ago
  Indexing doesn't replace the book, it's just a different way to read it. The index is useless without the book itself.
  - bonzini 3y ago
    Neither does knowing which parts of the book are darker or happier.
rgoulter 3y ago
IANAL. But, what's your consideration such that you think this usage of the books wouldn't be "fair use"? e.g. in the US https://en.wikipedia.org/wiki/Fair_use https://en.wikipedia.org/wiki/Fair_use factors of fair use are "purpose/character of the work", "nature of the copyrighted work", "amount/substantiality of the copyrighted work", "effect on the market for the copyrighted work". The website shows a few statistics computed from a book, and a few excerpts from the book. I'd think a consideration of those fair use factors favours the website: e.g. you're not going to look at those statistics/excerpts instead of reading the book. The website only shows a small portion of the book. The website's intention is to be educational.
- jprete 3y ago
  I think the site had a better claim by far than AI training does.
harshreality 3y ago
"Fair use" only applies to instances of copying / redistributing. The hint is in the name: copy-right. There's a notion, which seems to have taken off among creators who are paranoid about AI eating their livelihoods (which it might eat a chunk of) that copyright prevents people from doing anything with works they [legally] acquired other than personally read, listen, or watch it. That's not how copyright, as it has existed in the past, works. You can do all the algorithmic processing of your ebook collection that you want. You might be able to display small portions of a book to others, depending on the situation. Quoting one or two paragraphs out of an entire book seems like reasonably safe fair use, but that won't stop a copyright-maximalist creator (or their publisher) from suing you, and won't stop some copyright-maximalist judge from ruling against you, so it's probably best to minimize the amount of content from a book that you redisplay directly. But you can do all the analysis and statistics generation you want, and display those results to others. It remains to be seen what judges will do with AI generation of works based on ingesting gigantic amounts of copyrighted work. The entire framework of copyright is going to be broken, and until Congress steps in and changes it, judges are going to go every which way. There's no bright line for 4-factor analysis; it's always been a gut-level "is this a reasonable use that doesn't impact commercial sales too much". There's no possible rational way to draw a line. AI models can generate a painting of a new subject only loosely in the style of a contemporary painter, which would not be copyright infringement, or it can generate a near-clone of an existing work with the right prompting, and depending on how clever the prompter is, a lot of intermediate stages of likeness. Who decides how close to an existing work is too close?
- palata 3y ago
  Say I make a tool where you can enter the title of a book, and get the full text of the book without paying for it. I assume we all agree that would be illegal, right? Now say that instead of distributing that tool as an executable, I distributed it as a library. It would contain all the same books as the illegal executable above, but some developers would need to write an actual executable that would use the library. Obviously that executable would be illegal (same as the first one above), but would the library be illegal? I would think that we all reasonably agree that my library would be illegal, because I would essentially be selling the content of books for which I don't own the copyright. Now what happens when I distribute a trained model that has the potential to generate illegal work? Is it suddenly legal for me to sell it, because some developers write legal executables with it? How is that different from my library example above?
tensor 3y ago
There is no AI training here. Did you read the article?

This is an incredibly biased article, hinging entirely on the assumption that AI training is fair use.

by jprete 3y ago

This is an incredibly biased article, hinging entirely on the assumption that AI training is fair use.

Unfortunately it seems majority of HN clings on "AI is fair use!" the same way GIFs (of movie and TV snippets) are "perceived to be fair use" (but is simply not fair use in most cases).

542458 3y ago
Why do you feel that the Google vs Authors Guild ruling cited in the article isn’t relevant?
robertlagrant 3y ago
"majority" and "clings on" seem entirely arbitrary and perjorative. Why not make your actual point instead of just beating up a straw man?
luma 3y ago
And the other half of HN loses their mind when AI is brought up, completely missing the point made quite clearly in the OP that _no_AI_was_involved_in_any _of_this_. The AI hysteria on this site is ridiculous.
- quickthrower2 3y ago
  The ai moving goal! No 2023.ai was used. Maybe some 1993.ai. Chat GPT will be considered “just stats” in 2033.

pmarreck 3y ago

There is no “AI” in the project. It literally just seems to do Bayesian sentiment analysis of books. It does NOT then mine that data to score OTHER books. Please actually read the article next time before commenting

jprete 3y ago
I did in fact read the article. The author isn’t concerned about the book project. The author wants to use it as a springboard for complaining about AI fear, and his argument fundamentally hinges on the claim that AI training is fair use. He’s claiming the book project is comparable to AI training. I think the book project has a better claim to fair use than AI training does, although excerpting entire pages is maybe too far (but the word highlighting does feel like commentary).
- getabrainplz 3y ago
  If I read the book, take notes about the number of times $x happened and later post them to my website, are you fine with that? if I automate that process with a script that uses regular expressions, are you fine with that? if I add a neural network to my script and it spits out the same numbers, are you still fine with that? hopefully you see the madness in your argument.
  - jprete 3y ago
    I’m thinking specifically of generative AI’s fair use claims. I thought the book site (and all of your extension examples) was clearly fair use.

542458 3y ago

The article cites the Google vs Authors Guild case (https://www.techdirt.com/2013/11/14/google-gets-total-victory-over-authors-guild-book-scanning-is-fair-use/ https://www.techdirt.com/2013/11/14/google-gets-total-victor...) which was a total victory for Google. This seems fairly conclusive that the textual analysis here is fair use to me. > Similarly, Google Books is also transformative in the sense that it has transformed book text into data for purposes of substantive research, including data mining and text mining in new areas, thereby opening up new fields of research. Words in books are being used in a way they have not been used before. Google Books has created something new in the use of book text — the frequency of words and trends in their usage provide substantive information. Furthermore, is this actually AI training? This just looks like stats based on heuristics to me, I.e., garden variety sentiment analysis.

mycologos 3y ago
I think Google Books is cool, but "the frequency of words and trends in their usage provid[ing] substantive information" pre-dates Google Books by a long time. For example, there's a collection of word frequencies in the complete works of John Keats from 1917 [1]. Manually tabulated, too! [1] https://catalog.hathitrust.org/Record/001023999 https://catalog.hathitrust.org/Record/001023999
- 542458 3y ago
  Sure, but the Google Books case is a massive and well-funded court case that was a pretty resounding victory for Google and textual analysis of copyrighted works in general, so anybody arguing that this is obviously copyright violation needs to explain why the Google Books case isn’t relevant.

js8 3y ago

Why it wouldn't be? Indexing is fair use. I think where it gets murky with fair use would be if AI could actually plagiarize the book, but other than that, it should be fair use.

Nullabillity 3y ago
Indexing doesn't replace the book, it's just a different way to read it. The index is useless without the book itself.
- bonzini 3y ago
  Neither does knowing which parts of the book are darker or happier.

rgoulter 3y ago

IANAL. But, what's your consideration such that you think this usage of the books wouldn't be "fair use"? e.g. in the US https://en.wikipedia.org/wiki/Fair_use https://en.wikipedia.org/wiki/Fair_use factors of fair use are "purpose/character of the work", "nature of the copyrighted work", "amount/substantiality of the copyrighted work", "effect on the market for the copyrighted work". The website shows a few statistics computed from a book, and a few excerpts from the book. I'd think a consideration of those fair use factors favours the website: e.g. you're not going to look at those statistics/excerpts instead of reading the book. The website only shows a small portion of the book. The website's intention is to be educational.

jprete 3y ago
I think the site had a better claim by far than AI training does.

harshreality 3y ago

"Fair use" only applies to instances of copying / redistributing. The hint is in the name: copy-right. There's a notion, which seems to have taken off among creators who are paranoid about AI eating their livelihoods (which it might eat a chunk of) that copyright prevents people from doing anything with works they [legally] acquired other than personally read, listen, or watch it. That's not how copyright, as it has existed in the past, works. You can do all the algorithmic processing of your ebook collection that you want. You might be able to display small portions of a book to others, depending on the situation. Quoting one or two paragraphs out of an entire book seems like reasonably safe fair use, but that won't stop a copyright-maximalist creator (or their publisher) from suing you, and won't stop some copyright-maximalist judge from ruling against you, so it's probably best to minimize the amount of content from a book that you redisplay directly. But you can do all the analysis and statistics generation you want, and display those results to others. It remains to be seen what judges will do with AI generation of works based on ingesting gigantic amounts of copyrighted work. The entire framework of copyright is going to be broken, and until Congress steps in and changes it, judges are going to go every which way. There's no bright line for 4-factor analysis; it's always been a gut-level "is this a reasonable use that doesn't impact commercial sales too much". There's no possible rational way to draw a line. AI models can generate a painting of a new subject only loosely in the style of a contemporary painter, which would not be copyright infringement, or it can generate a near-clone of an existing work with the right prompting, and depending on how clever the prompter is, a lot of intermediate stages of likeness. Who decides how close to an existing work is too close?

palata 3y ago
Say I make a tool where you can enter the title of a book, and get the full text of the book without paying for it. I assume we all agree that would be illegal, right? Now say that instead of distributing that tool as an executable, I distributed it as a library. It would contain all the same books as the illegal executable above, but some developers would need to write an actual executable that would use the library. Obviously that executable would be illegal (same as the first one above), but would the library be illegal? I would think that we all reasonably agree that my library would be illegal, because I would essentially be selling the content of books for which I don't own the copyright. Now what happens when I distribute a trained model that has the potential to generate illegal work? Is it suddenly legal for me to sell it, because some developers write legal executables with it? How is that different from my library example above?

tensor 3y ago

There is no AI training here. Did you read the article?