Anti-Circumvention is Happening Again

In September of 2025, Anthropic settled a lawsuit, agreeing to pay $1.5 billion in fines to book publishers for using pirated copies of books to train their ai models.

Did you know piracy is actually a crime? Individuals have been fined, harrased, and even imprisoned for digital piracy. While there are some serious offenses that are arguably punishable, none of these people proved a serious risk to society at large or the individuals around them.

Gary Bowser was sentenced to 40 months in prison and ordered to pay $14.5 million in damages to Nintendo. Because he is unable to afford the damages, Nintendo collects 30% of his salary. Gary was involved with a group who designed and built hardware that would circumvent Nintendo devices and allow users to play unauthorized copies of games. ^{this is super script The Gaurdian}

Thomas Rasset https://en.wikipedia.org/wiki/CapitolRecords,Inc.v.Thomas-Rasset

Let's take a look at similar violations that various companies pursuing ai have committed:

It was revelaed that Meta literally used BitTorrent, a torrenting software that is used to download illegal copies of books, movies, music, and all forms of software, to mirror a popular pirating repository of books. Mark Zuckberg even knew and approved the operation which resulted in over 81 terabytes of books being pirated and used in their training datasets. ^{Ars Technica}

...

I'm not arguing that the people who are convicted of piracy shouldn't be convicted of crimes. This article is designed to note that these people have had their lives ruined for committing these crimes, compare that treatment to the treatment of the ai companies who are provably doing the exact same thing, but on a much larger, dangerous, and reckless scale, are not getting punished for it. Not a single person at these companies has been arrested or convicted.

These invidiaul people have been severely punished for circumventing technological incumbants. However when these companies do the same thing, they are rewarded with large sums of investment capital. The worst punishment they receive is a fine worth nothing more than a slap on the wrist.

But I don't want to stop there. Yes the world is currupt, the world is unfiar. But what would be a fiar punishment for these offenses? What would actually resolve the ill's these companies have inflicted on society? The real lost wages of artists? The lost compensation for these authors? The youtubers who have had their videos stolen?

I don't think these companies should be punished for trying to inovate and find new and novel technological achivements. This new iteration of tech companies has exposed the fact that to achieve LLM level tech, companies do need access to the entireity of the worlds knowledge base, that includes copywrighten works. I suggest we append the class of information that falls under copywright to allow for such innovation while respecting the authors who wrote and contributed to those data sources. How do we do this fairly?

Copywright law should be updated to mandate: If a company uses copywrighten content in their training data, wheather that be books, art works, images, videos, etc, that they must open source and publish the models, weights, and training architectures of any ai systems those works are based upon. Comercialization of ai systems trained on copywrighten data without compensation to authors is unethical. Therefore if a company is not able to compensate the authors of the data they are training their systems on, the systems should be made public so that those authors and the general public can benifit from them free of charge.

I find this approach far more compelling then slapping these companies with fines. In times past, companies like Microsoft were given fines for monopolizing the operating system and browser markets. That didn't hurt them in the long run. What actually made them bleed and led the way for the next generation of tech companies to arise was being mandated to open their operating system to allow for third party browsers like Netscape, and eventually Google. Google, a $4 trillion company (as of 2026), wouldn't exist if not for a law that mandated microsoft to open their prioprietary environment to third parties. A law mandating circumvention produced the greatest tool for information retreival up until that point.

This approach isn't perfect, yes the creators, authors, and artists are still not getting direct compensation for their works. The companies will still benifit from their transgressions. You could even argue that users of these models get access, perhaps indirectly, to the copywrighten content of the authors works without compensation either. All fair and valid. I'm not certain if my proposal solves this problem. But what it does solve is actually punishing these companies for breaking the law. Exposing their internal architectures, model weights, and binaries to their competitors will damage their moat.

The unfortunate truth that doesn't have an easy solution is simply: the toothpaste has been squeezed from the tube.

https://news.bloomberglaw.com/ip-law/openai-risks-billions-as-court-weighs-privilege-in-copyright-row

https://www.theguardian.com/technology/2025/sep/05/anthropic-settlement-ai-book-lawsuit

https://www.theguardian.com/technology/2025/jan/10/mark-zuckerberg-meta-books-ai-models-sarah-silverman

https://www.yahoo.com/news/meta-faces-lawsuit-training-ai-155056631.html

https://www.google.com/search?q=what+do+pirates+think+of+ai+companies+stealing+stolen+books&client=ubuntu-sn&hs=OGL&scaesv=846ebd86ebd88482&channel=fs&sxsrf=ANbL-n73ztEOPGP8V0AUo2Kg1NO4faT5-g%3A1771951869777&ei=dadaemXL9TLp84PoPXIuQI&biw=2503&bih=1528&ved=0ahUKEwipioeOyKSAxXU5ckDHaA6MicQ4dUDCBE&uact=5&oq=what+do+pirates+think+of+ai+companies+stealing+stolen+books&gslp=Egxnd3Mtd2l6LXNlcnAiO3doYXQgZG8gcGlyYXRlcyB0aGluayBvZiBhaSBjb21wYW5pZXMgc3RlYWxpbmcgc3RvbGVuIGJvb2tzMgQQABhHMgQQABhHMgQQABhHMgQQABhHMgQQABhHMgQQABhHMgQQABhHMgQQABhHSM8BUABYAHAAeAKQAQCYAQCgAQCqAQC4AQPIAQCYAgGgAgaYAwDiAwUSATEgQIgGAZAGCJIHATGgBwCyBwC4BwDCBwMyLTHIBwSACAE&sclient=gws-wiz-serp