How Aaron Swartz Built AI (Literally)

(this opinion essay was written in 2024)

Aaron Swartz during PIPA Rally

In the cramped wiring closet of MIT’s Building 16, Aaron Swartz made a bold move that altered the course of public’s understanding about open-source data. He was looking for a way to break the locks on human knowledge. Fast forward to 2024, and the “**AI Race**” between global superpowers has turned that same knowledge into the most valuable commodity on earth. Oftentimes, China and USA flex their might in technology. But, their AI Models , short term for Artificial Intelligence models, are hungry for something that the open web can’t reliably provide: the truth. In order to build such intelligent system, they need peer-reviewed, evidence-based research that Aaron once risked everything to set free.

Sometime in the past, there was a person often dubbed “The Internet’s Own Boy,” Aaron Swartz, who tried to democratize research journals for everyone to see. Spoiler alert: he released an immense amount of these journals by hacking servers, entering a restricted basement wiring closet in MIT’s Building 16 (Room 16–004T), and physically hard-wiring a laptop into the network. He would transport his external drives from that closet back to his hideouts, uploading millions of articles to the internet. Silicon Valley and the government eventually drove him to his demise. This noble action eventually made information freely available. But what is infuriating to me, and everyone else off, is that if corporations need something, they just take it. These AI models benefited from his sacrifice. The nobility and sacrifice of Aaron’s action was rooted in the idea that knowledge belongs to humanity. By using his work to build chatbots and AI products, these companies have effectively stolen a gift meant for the people and sold it back to them.

[embed]

I stumbled upon his name during the pandemic, in 2020, when I watched this documentary film called “Internet’s Own Boy”. I learned that at such a young age, he was one of the founders of social media site named Reddit. He is the quintessential intelligent man with integrity. In 2008, he wrote and published the Guerilla Open Access Manifesto” where he argued that “information is power” and it is a moral imperative rather than corporations implementing paywalls just to read them. He argued that forcing people to pay the work of their collegues, especially if the research was often made possible by the taxpayers, was a “private theft if public culture”. I felt so connected to his story that I even made a wireframe website dedicated to him as a school project on Wix. I was trying to spread his message as soon as I learned about it.

[embed]Internet’s Own Boy

The US government’s response was deliberate and brutal. In 2011, he was charged with 14 counts of felony charges under the country’s Computer Fraud and Abuse Act. Even after Aaron returned the data and JSTOR wasn’t keen on going after the hacktivist, the government refused to back down. He was indicted with 35 years in prison and ordered to pay damages of a million dollars. This pressure and this thought of him being a felon for an act of civil disobedience led Aaron to take his own life on January 11th, 2013, at just 26 years old.

His tragedy is a perfect display of the asymmetry of justice. Aaron Swartz was hounded to his death after liberating data that he intended to share with the humanity. Fast forward to the AI boom, and we see corporation scaping vast amount of data from these open access initiative to…. build a billion-dollar product? This is where the problem set to begin. Aaron was treated as if he’s a terrorist for using a script to download huge number of papers. Today, AI companies use even more sophisticated and advanced “Scrapers” to take everything on the web and yet, they were even called as “innovators”.

The very datasets that companies like OpenAI, Meta, and Google now use were created by Aaron’s fight for Open Access. Many AI models are trained on massive datasets like “The Pile” or “Books3.” Investigations and court filings, such as the 2024 Meta lawsuits, have revealed that these datasets contain millions of documents and research papers from “shadow libraries” like LibGen and Sci-Hub, which are the digital descendants of Aaron’s work, and running it through their algorithms to scrub away its “stolen” origin. By the time it reaches the consumer as a chatbot response, the sacrifice is gone, replaced by a subscription fee.

This brings us to Alexandra Elbakyan, which is the founder of Sci-Hub. Often called the “spiritual successor” to Aaron Swartz, Elbakyan, a programmer from Kazakhstan, built the very thing Aaron was trying to: a platform that hosts over 80 million research papers for free. Even though she lives in exile to avoid the same fate as Aaron, her work provided the “clean” data that AI researchers needed to make their models smart.

Imagine the absurdity of it: knowledge like Darwin’s theory of evolution, which is a scientific foundation that belong to all of humanity, is locked behind… a paywall. Why do you have to pay to access his documents? Where does that money go? To his grave? The afterlife? No, it goes to a selected few who benefit off the knowledge of people. Aaron wanted this data free so a student in Africa could learn. So that a man like me from my country, the Philippines, a third-world nation riddled with education insecurity, could learn about the world beyond his knowledge. Instead, AI companies have “vacuumed” it up to create a “black box.” AI companies have “vacuumed” this data to sharpen machines that are increasingly used to replace those very humans. They unlock the library so their software can “read” it for us and after, charge us for the summary.

They called Aaron a criminal for wanting us to know too much. Today, they call AI “revolutionary” for knowing everything. Somewhere in the billions of parameters of these language models, Aaron’s liberated journals are still there, and whispering truths about medicine, physics, and justice. But the machine doesn’t care about all these “moral imperative” it only cares about the next token, about the next thing you will ask and go about your day.

As I look at my school project on Wix, dedicated to a man I never met, I realize that the “AI Race” is missing the one thing Aaron had in abundance: which is a human soul. We are currently teaching machines how to speak, but we are forgetting how to hear the voices of those who broke their lives to give us the truth. Aaron Swartz didn’t die off the face of the earth so that a computer could summarize Darwin for a fee barely anyone can afford nowadays. He died so that the world could think for itself. If we want to truly honor the Internet’s Own Boy, we must ensure that the “intelligence” of the future is not a product that we can just buy with your money, but a fire we refuse to let them extinguish. The library is already open, we just have to decide if we’re brave enough to walk past the toll booth and claim what was always ours.