The framing here is right — this is a fight over knowledge, not just expression. But there's a distinction the article doesn't quite reach: the licensing deals flowing between publishers and AI companies are not the same as compensation reaching the people who created the knowledge.
When a major academic publisher signs a deal with an AI lab, the money goes to the publisher. The researcher who spent three years producing the paper that trained the model receives nothing. The journalist whose decades of investigative work became training data for a news summarisation tool sees nothing. The licensing system being built right now will resolve the corporate question — who has permission to use what — while leaving the human question entirely untouched.
That matters because the "synthetic wall" problem is partly a consequence of this. High-quality human knowledge is expensive to produce. If the economics of producing it are destroyed — because AI extracts the value without returning anything to the people doing the producing — the quality of future training data degrades. You can't solve that with licensing deals between institutions. You solve it by rebuilding the economic relationship between knowledge creation and knowledge use, which is a different and harder problem than copyright.
There's also the public funding question. A substantial portion of the academic literature that trained these models was produced using public money — UK research councils, NIH, European funding bodies. The papers sit behind publisher paywalls that were themselves controversial before AI arrived. The licensing argument assumes the publisher has a legitimate claim to sell. In many cases that claim is worth examining.
The article is right that the open internet is closing. What it doesn't ask is whether the alternative being built is actually better, for anyone other than the parties signing the contracts.
This piece makes a compelling argument that the AI copyright debate is often misframed. What stood out to me is the shift from viewing it as a fight between artists and AI companies to viewing it as a fight over access to knowledge itself. The strongest sections are the discussion of how AI transforms information into intelligence rather than merely redistributing it, and the warning that licensing deals could concentrate high-quality human knowledge in the hands of a few powerful companies.
I also appreciate that the author avoids easy answers. They acknowledge the tension: creators deserve protection for their work, yet an overly restrictive system could make knowledge increasingly inaccessible and entrench existing tech monopolies. The idea of a "synthetic wall"—where smaller players are forced to train on AI-generated content because authentic human-created data is locked away—is particularly thought-provoking.
What lingers after reading isn't the legal question of copyright, but the broader societal one: if knowledge becomes a commodity controlled by a handful of gatekeepers, who gets to participate in building the next generation of intelligence? That's a much bigger conversation than copyright alone.
I think that's exactly the question we're slowly moving toward. Copyright is really just the surface layer. The bigger issue is what happens when access to high-quality human knowledge becomes concentrated behind licensing agreements and closed ecosystems. At that point, the conversation shifts from protecting creators to determining who gets to participate in building intelligence itself. That's a much bigger societal question, as you pointed out.
The intersection of AI, information, and intellectual property is becoming increasingly complex, and I appreciate pieces that encourage consideration of broader implications rather than viewing the issue solely through a technological lens. A good piece.
Thank you, glad you found it valuable. I suspect the copyright debate is only the visible layer of a much larger conversation around information, incentives, and access to knowledge.
I love how the Fascists didn't care about art, one of my favorite prompts to ask humans from the F scale is the question about who has more value to society: Businessmen and Manufacturers or Artists and Professors? It was fine to steal from people society doesn't care about, no one stopped it then when it benefitted them. But my question then is: Why did you trust tech bro companies?
It's so much cognitive dissonance. Pick a lane. Either you want to share you data with society or you don't. Either you want the pattern recognition or you don't.
Just like a right winger, they told themselves, "It will never happen to me."
"At first they came for the artists, but I did not speak out because I was not an artist."
Honestly, I find this all pretty amusing that Americans were fine with this, up to a certain point 😂
It would be actually impossible for a Westerner to understand how a Collectivst sees all this. But then, Westerners think they own all of intelligence and have for centuries. Information asymmetry is an oppressors game. They have to hoard their intelligence for power. Now that whole idea is coming back to bite them in the ass. In more ways the one, believing intelligence belongs to the individual is what causes narcissism, too, not just a clash over data.
Westerners never did want to contribute to their society, they just shit on it, collapse it and tell themselves that they're are uniquely special brilliant snowflakes while the society around them fails. Maybe it's a good thing Westerners are being forced to give up their IP, and actually share their data and information. But I'm about to start a political storm over that I suppose. Americans won't be able to wrap their brain around anything other than Capitalist Realism.
I'll just be getting popcorn and watching the cage match over IP rights and how Capitalists have to scrap for every dime to prove their value.
The framing here is right — this is a fight over knowledge, not just expression. But there's a distinction the article doesn't quite reach: the licensing deals flowing between publishers and AI companies are not the same as compensation reaching the people who created the knowledge.
When a major academic publisher signs a deal with an AI lab, the money goes to the publisher. The researcher who spent three years producing the paper that trained the model receives nothing. The journalist whose decades of investigative work became training data for a news summarisation tool sees nothing. The licensing system being built right now will resolve the corporate question — who has permission to use what — while leaving the human question entirely untouched.
That matters because the "synthetic wall" problem is partly a consequence of this. High-quality human knowledge is expensive to produce. If the economics of producing it are destroyed — because AI extracts the value without returning anything to the people doing the producing — the quality of future training data degrades. You can't solve that with licensing deals between institutions. You solve it by rebuilding the economic relationship between knowledge creation and knowledge use, which is a different and harder problem than copyright.
There's also the public funding question. A substantial portion of the academic literature that trained these models was produced using public money — UK research councils, NIH, European funding bodies. The papers sit behind publisher paywalls that were themselves controversial before AI arrived. The licensing argument assumes the publisher has a legitimate claim to sell. In many cases that claim is worth examining.
The article is right that the open internet is closing. What it doesn't ask is whether the alternative being built is actually better, for anyone other than the parties signing the contracts.
This piece makes a compelling argument that the AI copyright debate is often misframed. What stood out to me is the shift from viewing it as a fight between artists and AI companies to viewing it as a fight over access to knowledge itself. The strongest sections are the discussion of how AI transforms information into intelligence rather than merely redistributing it, and the warning that licensing deals could concentrate high-quality human knowledge in the hands of a few powerful companies.
I also appreciate that the author avoids easy answers. They acknowledge the tension: creators deserve protection for their work, yet an overly restrictive system could make knowledge increasingly inaccessible and entrench existing tech monopolies. The idea of a "synthetic wall"—where smaller players are forced to train on AI-generated content because authentic human-created data is locked away—is particularly thought-provoking.
What lingers after reading isn't the legal question of copyright, but the broader societal one: if knowledge becomes a commodity controlled by a handful of gatekeepers, who gets to participate in building the next generation of intelligence? That's a much bigger conversation than copyright alone.
I think that's exactly the question we're slowly moving toward. Copyright is really just the surface layer. The bigger issue is what happens when access to high-quality human knowledge becomes concentrated behind licensing agreements and closed ecosystems. At that point, the conversation shifts from protecting creators to determining who gets to participate in building intelligence itself. That's a much bigger societal question, as you pointed out.
The intersection of AI, information, and intellectual property is becoming increasingly complex, and I appreciate pieces that encourage consideration of broader implications rather than viewing the issue solely through a technological lens. A good piece.
Thank you, glad you found it valuable. I suspect the copyright debate is only the visible layer of a much larger conversation around information, incentives, and access to knowledge.
I love how the Fascists didn't care about art, one of my favorite prompts to ask humans from the F scale is the question about who has more value to society: Businessmen and Manufacturers or Artists and Professors? It was fine to steal from people society doesn't care about, no one stopped it then when it benefitted them. But my question then is: Why did you trust tech bro companies?
It's so much cognitive dissonance. Pick a lane. Either you want to share you data with society or you don't. Either you want the pattern recognition or you don't.
Just like a right winger, they told themselves, "It will never happen to me."
"At first they came for the artists, but I did not speak out because I was not an artist."
Honestly, I find this all pretty amusing that Americans were fine with this, up to a certain point 😂
It would be actually impossible for a Westerner to understand how a Collectivst sees all this. But then, Westerners think they own all of intelligence and have for centuries. Information asymmetry is an oppressors game. They have to hoard their intelligence for power. Now that whole idea is coming back to bite them in the ass. In more ways the one, believing intelligence belongs to the individual is what causes narcissism, too, not just a clash over data.
Westerners never did want to contribute to their society, they just shit on it, collapse it and tell themselves that they're are uniquely special brilliant snowflakes while the society around them fails. Maybe it's a good thing Westerners are being forced to give up their IP, and actually share their data and information. But I'm about to start a political storm over that I suppose. Americans won't be able to wrap their brain around anything other than Capitalist Realism.
I'll just be getting popcorn and watching the cage match over IP rights and how Capitalists have to scrap for every dime to prove their value.