5 ms·

Outside of tech, AI isn't making any friends. It's an unexpected threat on the horizon for the bearers of accumulated privilege, such as monopolies and nation-s

Outside of tech, AI isn't making any friends. It's an unexpected threat on the horizon for the bearers of accumulated privilege, such as monopolies and nation-state autocrats. They hear echoes of the computer and internet revolutions, which were key in upending many prior "garden patches" of power. Their cronies have been stirring up resistance among workers on the front lines of disruption, such as some content creators. The truth is that no one knows where the dust will settle. All covered wagons are venturing west of the Rockies at the same time. Content creators are not really facing a worse outlook than everyone else. They are, however, in a better position to hinder the advance. In principle, an AI learning from a scientific textbook is no different than a human student doing the same. Neither will violate copyright law when they're done learning, except perhaps accidentally - paraphrasing and facts are not violations. Unfortunately, legal and ethical principles can differ from legal reality. We're left hoping that some altruistic legal minds will open up a Northwest Passage for us, like Thomas Penfield Jackson in US v. Microsoft. The worst possible outcome is that we end up with an all-powerful AI cartel which negotiates massive deals with IP conglomerates, locking out competition and open and free alternatives.

autoexec 2y ago
> In principle, an AI learning from a scientific textbook is no different than a human student doing the same. Neither will violate copyright law when they're done learning If AI were only learning from scientific textbooks very few people would have problem with it. When AI learns from creative works like novels, it runs into the same problems we have as humans. We can read a book, but the minute we start distributing copies of it from memory we're in legal trouble.
atlasunshrugged 2y ago
Why wouldn't nation state autocrats be interested in AI? Seems like a great tool for facial recognition at scale and eventually maybe even automated policing via drones and the like, thus sparing them the costs (and danger) of having to maintain a major police/military force to protect their regimes
- worstspotgain 2y ago
  They've largely already mastered control of their domains, e.g. by outlawing all non-state media. Generally speaking, if you're already at the top, everything is mostly a threat simply because nothing is much of an opportunity.
- whodidntante 2y ago
  Yes, nation states will be interested, and it goes far beyond simple tools like automated policing. Once AI is used as the buffer between all forms of communication (making sure the emails I send put me in the best light, filtering incoming email to make me more productive, generating web content, reading web content in a way that makes me more efficient, even voice and video calls, the nation state can not only monitor, but also subtly alter all communications, in real time, and in context to guide/enforce how people interact/communicate and think. This is already happening in areas such as search, getting "answers", and content generation. Your information is constrained/mapped into what the the AI deems appropriate. We already know that the NSA taps into the communication backbone of the internet. Why not add a real-time AI monitor/filter to best manage how we all communicate ?
zifpanachr23 2y ago
This is a really peak crazy libertarian take on the issue of artists and content creators being peeved that AI is making crappier content for cheaper and driving down their job opportunities and wages, after training on their content (and sometimes regurgitating it word for word in large chunks) and not paying royalties. If you think artistic people of various kinds are doing particularly well and need to have their privilege checked to stop the discrimination against AI tech bros...then I think you are out of touch. The starving artist meme is a thing for a reason, and most content creators are not making a lot of money. They enjoy it, and maybe make a little on the side. But most work a shitty day job as well to fund their passion, the thing that makes them enjoy life and helps them find meaning. They just want to contribute to culture and enjoy making art but wouldn't be able to justify it if all the existing low wage starving artist work dries up in favor of essentially free AI generated crap. I'd hazard a guess that the majority of the country would take their side and call Silicon Valley tech bros and VCs privileged and anti-social. I think a culture in which it is non-viable for an ordinary person to participate in art and culture on the side is kind of sad. I value these things because there is a person on the other end and it helps me empathize and try and understand how they are feeling and thinking. An AI kills that. You'll get great stock photos generic advertising, but you'll destroy something fundamental to human culture.
- cherryteastain 2y ago
  Everything not expressly prohibited by law is permitted. Extending copyright retroactively to cover AI training is like prohibiting the wearing of blue shorts on Thursdays and fining anyone who has ever worn blue shorts on a Thursday.
- worstspotgain 2y ago
  If my college textbook's author demanded a percentage cut of my future earnings I'd be much more peeved than content creators are now. I think pointing out that it's crappy is a much better critique of crappy content than looking at how it was generated. As for tech bros vs. artists, that's a reality show I haven't watched and one I'm not interested in. I'm interested in progress and human advancement, and I'm definitely not a Libertarian or tech bro.
unyttigfjelltol 2y ago
> In principle, an AI learning from a scientific textbook is no different than a human student doing the same. In principle human brains don't count as a copy of a copyrighted work, whereas the AI model that is capable of regurgitating the original-- I don't understand how that's not a copy[1] or derivative work which is a good indication the judges presented with these questions won't understand either. So, no, a brain with knowledge is a brain, and an AI model with knowledge is a potentially infringing copy or derivative work. [1] https://www.law.cornell.edu/definitions/uscode.php?width=840&height=800&iframe=true&def_id=17-USC-2024104691-364936160&term_occur=999&term_src= https://www.law.cornell.edu/definitions/uscode.php?width=840...
- worstspotgain 2y ago
  I can memorize and recite a poem, but in the context of a scientific textbook, neither I nor an AI will remember the actual words used, just key facts like E = mc^2. If you limit your critique to the AI's ability to recite a poem or lyrics verbatim, then you have a point. However, I used the example of a scientific textbook for that exact reason. To the extent that the AI only paraphrases or cites facts from, it's not a copy. The derivative work issue is more interesting. The AI can generate a derivative work, but won't do so without a prompt asking for one. If you ask for an Oz-like bedtime story, you should get an Anxious Lynx and not a Cowardly Lion. If you ask for a story with Oz characters, the result will be a derivative work, but it will include your prompt. The work was not in the AI, though the ability to make it was. In this sense, the AI is more like an instrument that can be used to violate copyright law, like an employee who has read Oz would be. It's hard to find a close inanimate analogue, but it's probably on the same shelf as search engines and typewriters.
  - dahart 2y ago
    > neither I nor an AI will remember the actual words used […] To the extent that the AI only paraphrases Many AIs have in fact reproduced copyrighted work in large portions, it’s one of the reasons there are a bunch of lawsuits. > AI can generate a derivative work, but won’t do so without a prompt asking for one This is an odd and vague claim. Which AI are you talking about? And why do you claim there’s any generative AI in existence that won’t remix from it’s training examples? Their design from the ground up is to create derivative works. > AI is more like an instrument that can be used to violate copyright law If the training data wasn’t legally acquired, using it to train in the first place may violate the law. People are trying to change copyright’s interpretation, but pending the outcome of that it’s pretty clear that today’s AI may have violated copyright law before generating any output.
evdubs 2y ago
> In principle, an AI learning from a scientific textbook is no different than a human student doing the same. It's not about the process of building a model. It's about the copyright infringement to acquire the training set and it's about the copyright infringement when producing output. Humans, just like these AI companies, can infringe copyrights by downloading copyrighted material. Humans, just like AI programs, can produce material that is derivative of material protected by copyright. Whatever the model represents and whatever is in your brain is not high on the list of concerns.
- worstspotgain 2y ago
  I've addressed the "when producing output" part in another reply, so here are my thoughts on the "to acquire the training set" part. Unless you're downloading from something like Z-Library (which incidentally is a deliberate attack on Western IP on Putin's behalf) you generally have the fair use rights to "digest" the information you receive. You can process the contents of a webpage or momentarily OCR a printed page, provided that it was legally distributed and that you don't store a copy proper. There can be specialized restrictions if it's something that you have a license for instead of a copy of, see e.g. https://en.wikipedia.org/wiki/Shrinkwrap_(contract_law) https://en.wikipedia.org/wiki/Shrinkwrap_(contract_law) As for derivative works, not all the works that are made "in proximity of" previous works are derivative works. You can make a simulated reality movie without infringing on the Matrix, for instance. Just write a somewhat different plot and don't name your lead character Neo.
  - evdubs 2y ago
    > You can process the contents of a webpage or momentarily OCR a printed page, provided that it was legally distributed and that you don't store a copy proper. The AI companies are definitely storing proper copies of their "corpus". > You can make a simulated reality movie without infringing on the Matrix, for instance. You can also make a simulated reality movie, even without the name Neo, and a court could find it substantially similar to the Matrix and thus copyright infringing. It doesn't need to be the same.
JKCalhoun 2y ago
AI will train from AI then. The copyright source is twice removed.
forgetfulness 2y ago
> Outside of tech, AI isn't making any friends. It's an unexpected threat on the horizon for the bearers of accumulated privilege, such as monopolies and nation-state autocrats How so? Generative AI has developed in a way that's very amenable to the bearers of accumulated privilege, it requires so much capital expenditure that it's in the hands of a few companies that can be variously negotiated with or strong armed. This is not at all like the computer and internet revolutions of the 90s. > The worst possible outcome is that we end up with an all-powerful AI cartel which negotiates massive deals with IP conglomerates, locking out competition and open and free alternatives. What's the difference with the status quo if the main players are Microsoft, Meta and Google?
- worstspotgain 2y ago
  > How so? Generative AI has developed in a way that's very amenable to the bearers of accumulated privilege, it requires so much capital expenditure that it's in the hands of a few companies that can be variously negotiated with or strong armed. This is not at all like the computer and internet revolutions of the 90s. Even at today's bloated GPU prices (which won't last,) training rigs cost about the same as the servers and data pipes that serious dot-coms needed, and much less than the fabs that serious semiconductor outfits needed in ~1980, especially when adjusted for inflation. Many AI startups don't do their own primary training at all, making their costs identical to barebones dot-coms'. > What's the difference with the status quo if the main players are Microsoft, Meta and Google? You forgot OpenAI, Anthropic, Apple and Amazon. Of the seven, two are new, three emerged from the last revolution and two from the one before. The incumbent companies in the 90s were the likes of IBM, DEC, HP, Unisys, SGI and Sun. Almost all of them smelled the coffee and invested heavily in various Internet plays. Only a few survived, most of them only barely. The previous revolution was even bloodier. Today's incumbents are also smelling the coffee and have AI plays. None of them is certain they'll still be around 10 years from now. The reason Meta opened up Llama is because they were convinced that the barriers to entry were minimal. In 2022, Altman famously claimed that if there was a Moore's Law in parameter count, it had already topped out. A large fraction of use cases can be handled with models under 10B. Training algorithms are increasing in efficiency faster than parameter counts are increasing. The only real barriers to entry are patents (and we won't know what that looks like until they break stealth) and the possibility of nefarious anticompetitive deals with the IP conglomerates that we're discussing here. Many economists expect the disruption caused by AI to be much greater than that of PCs and the Internet. The usual parallel is electrification. The gentry from back East is in the covered wagons too.

autoexec 2y ago

> In principle, an AI learning from a scientific textbook is no different than a human student doing the same. Neither will violate copyright law when they're done learning If AI were only learning from scientific textbooks very few people would have problem with it. When AI learns from creative works like novels, it runs into the same problems we have as humans. We can read a book, but the minute we start distributing copies of it from memory we're in legal trouble.

atlasunshrugged 2y ago

Why wouldn't nation state autocrats be interested in AI? Seems like a great tool for facial recognition at scale and eventually maybe even automated policing via drones and the like, thus sparing them the costs (and danger) of having to maintain a major police/military force to protect their regimes

worstspotgain 2y ago
They've largely already mastered control of their domains, e.g. by outlawing all non-state media. Generally speaking, if you're already at the top, everything is mostly a threat simply because nothing is much of an opportunity.
whodidntante 2y ago
Yes, nation states will be interested, and it goes far beyond simple tools like automated policing. Once AI is used as the buffer between all forms of communication (making sure the emails I send put me in the best light, filtering incoming email to make me more productive, generating web content, reading web content in a way that makes me more efficient, even voice and video calls, the nation state can not only monitor, but also subtly alter all communications, in real time, and in context to guide/enforce how people interact/communicate and think. This is already happening in areas such as search, getting "answers", and content generation. Your information is constrained/mapped into what the the AI deems appropriate. We already know that the NSA taps into the communication backbone of the internet. Why not add a real-time AI monitor/filter to best manage how we all communicate ?

zifpanachr23 2y ago

This is a really peak crazy libertarian take on the issue of artists and content creators being peeved that AI is making crappier content for cheaper and driving down their job opportunities and wages, after training on their content (and sometimes regurgitating it word for word in large chunks) and not paying royalties. If you think artistic people of various kinds are doing particularly well and need to have their privilege checked to stop the discrimination against AI tech bros...then I think you are out of touch. The starving artist meme is a thing for a reason, and most content creators are not making a lot of money. They enjoy it, and maybe make a little on the side. But most work a shitty day job as well to fund their passion, the thing that makes them enjoy life and helps them find meaning. They just want to contribute to culture and enjoy making art but wouldn't be able to justify it if all the existing low wage starving artist work dries up in favor of essentially free AI generated crap. I'd hazard a guess that the majority of the country would take their side and call Silicon Valley tech bros and VCs privileged and anti-social. I think a culture in which it is non-viable for an ordinary person to participate in art and culture on the side is kind of sad. I value these things because there is a person on the other end and it helps me empathize and try and understand how they are feeling and thinking. An AI kills that. You'll get great stock photos generic advertising, but you'll destroy something fundamental to human culture.

cherryteastain 2y ago
Everything not expressly prohibited by law is permitted. Extending copyright retroactively to cover AI training is like prohibiting the wearing of blue shorts on Thursdays and fining anyone who has ever worn blue shorts on a Thursday.
worstspotgain 2y ago
If my college textbook's author demanded a percentage cut of my future earnings I'd be much more peeved than content creators are now. I think pointing out that it's crappy is a much better critique of crappy content than looking at how it was generated. As for tech bros vs. artists, that's a reality show I haven't watched and one I'm not interested in. I'm interested in progress and human advancement, and I'm definitely not a Libertarian or tech bro.

unyttigfjelltol 2y ago

> In principle, an AI learning from a scientific textbook is no different than a human student doing the same. In principle human brains don't count as a copy of a copyrighted work, whereas the AI model that is capable of regurgitating the original-- I don't understand how that's not a copy[1] or derivative work which is a good indication the judges presented with these questions won't understand either. So, no, a brain with knowledge is a brain, and an AI model with knowledge is a potentially infringing copy or derivative work. [1] https://www.law.cornell.edu/definitions/uscode.php?width=840&height=800&iframe=true&def_id=17-USC-2024104691-364936160&term_occur=999&term_src= https://www.law.cornell.edu/definitions/uscode.php?width=840...

worstspotgain 2y ago
I can memorize and recite a poem, but in the context of a scientific textbook, neither I nor an AI will remember the actual words used, just key facts like E = mc^2. If you limit your critique to the AI's ability to recite a poem or lyrics verbatim, then you have a point. However, I used the example of a scientific textbook for that exact reason. To the extent that the AI only paraphrases or cites facts from, it's not a copy. The derivative work issue is more interesting. The AI can generate a derivative work, but won't do so without a prompt asking for one. If you ask for an Oz-like bedtime story, you should get an Anxious Lynx and not a Cowardly Lion. If you ask for a story with Oz characters, the result will be a derivative work, but it will include your prompt. The work was not in the AI, though the ability to make it was. In this sense, the AI is more like an instrument that can be used to violate copyright law, like an employee who has read Oz would be. It's hard to find a close inanimate analogue, but it's probably on the same shelf as search engines and typewriters.
- dahart 2y ago
  > neither I nor an AI will remember the actual words used […] To the extent that the AI only paraphrases Many AIs have in fact reproduced copyrighted work in large portions, it’s one of the reasons there are a bunch of lawsuits. > AI can generate a derivative work, but won’t do so without a prompt asking for one This is an odd and vague claim. Which AI are you talking about? And why do you claim there’s any generative AI in existence that won’t remix from it’s training examples? Their design from the ground up is to create derivative works. > AI is more like an instrument that can be used to violate copyright law If the training data wasn’t legally acquired, using it to train in the first place may violate the law. People are trying to change copyright’s interpretation, but pending the outcome of that it’s pretty clear that today’s AI may have violated copyright law before generating any output.

evdubs 2y ago

> In principle, an AI learning from a scientific textbook is no different than a human student doing the same. It's not about the process of building a model. It's about the copyright infringement to acquire the training set and it's about the copyright infringement when producing output. Humans, just like these AI companies, can infringe copyrights by downloading copyrighted material. Humans, just like AI programs, can produce material that is derivative of material protected by copyright. Whatever the model represents and whatever is in your brain is not high on the list of concerns.

worstspotgain 2y ago
I've addressed the "when producing output" part in another reply, so here are my thoughts on the "to acquire the training set" part. Unless you're downloading from something like Z-Library (which incidentally is a deliberate attack on Western IP on Putin's behalf) you generally have the fair use rights to "digest" the information you receive. You can process the contents of a webpage or momentarily OCR a printed page, provided that it was legally distributed and that you don't store a copy proper. There can be specialized restrictions if it's something that you have a license for instead of a copy of, see e.g. https://en.wikipedia.org/wiki/Shrinkwrap_(contract_law) https://en.wikipedia.org/wiki/Shrinkwrap_(contract_law) As for derivative works, not all the works that are made "in proximity of" previous works are derivative works. You can make a simulated reality movie without infringing on the Matrix, for instance. Just write a somewhat different plot and don't name your lead character Neo.
- evdubs 2y ago
  > You can process the contents of a webpage or momentarily OCR a printed page, provided that it was legally distributed and that you don't store a copy proper. The AI companies are definitely storing proper copies of their "corpus". > You can make a simulated reality movie without infringing on the Matrix, for instance. You can also make a simulated reality movie, even without the name Neo, and a court could find it substantially similar to the Matrix and thus copyright infringing. It doesn't need to be the same.

JKCalhoun 2y ago

AI will train from AI then. The copyright source is twice removed.

forgetfulness 2y ago

> Outside of tech, AI isn't making any friends. It's an unexpected threat on the horizon for the bearers of accumulated privilege, such as monopolies and nation-state autocrats How so? Generative AI has developed in a way that's very amenable to the bearers of accumulated privilege, it requires so much capital expenditure that it's in the hands of a few companies that can be variously negotiated with or strong armed. This is not at all like the computer and internet revolutions of the 90s. > The worst possible outcome is that we end up with an all-powerful AI cartel which negotiates massive deals with IP conglomerates, locking out competition and open and free alternatives. What's the difference with the status quo if the main players are Microsoft, Meta and Google?

worstspotgain 2y ago
> How so? Generative AI has developed in a way that's very amenable to the bearers of accumulated privilege, it requires so much capital expenditure that it's in the hands of a few companies that can be variously negotiated with or strong armed. This is not at all like the computer and internet revolutions of the 90s. Even at today's bloated GPU prices (which won't last,) training rigs cost about the same as the servers and data pipes that serious dot-coms needed, and much less than the fabs that serious semiconductor outfits needed in ~1980, especially when adjusted for inflation. Many AI startups don't do their own primary training at all, making their costs identical to barebones dot-coms'. > What's the difference with the status quo if the main players are Microsoft, Meta and Google? You forgot OpenAI, Anthropic, Apple and Amazon. Of the seven, two are new, three emerged from the last revolution and two from the one before. The incumbent companies in the 90s were the likes of IBM, DEC, HP, Unisys, SGI and Sun. Almost all of them smelled the coffee and invested heavily in various Internet plays. Only a few survived, most of them only barely. The previous revolution was even bloodier. Today's incumbents are also smelling the coffee and have AI plays. None of them is certain they'll still be around 10 years from now. The reason Meta opened up Llama is because they were convinced that the barriers to entry were minimal. In 2022, Altman famously claimed that if there was a Moore's Law in parameter count, it had already topped out. A large fraction of use cases can be handled with models under 10B. Training algorithms are increasing in efficiency faster than parameter counts are increasing. The only real barriers to entry are patents (and we won't know what that looks like until they break stealth) and the possibility of nefarious anticompetitive deals with the IP conglomerates that we're discussing here. Many economists expect the disruption caused by AI to be much greater than that of PCs and the Internet. The usual parallel is electrification. The gentry from back East is in the covered wagons too.