You’ve felt it. That quiet dread when you watch a researcher feed a decades-old problem into an AI model and get an answer in seconds. Something’s wrong. Science isn’t supposed to be this easy. The struggle, the late nights, the accidental discoveries—they’re not bugs. They’re the whole point.
Now a group of well-meaning academics has released the Leiden Declaration, a set of ethical guidelines for using AI in scientific research. It sounds noble. It sounds like a shield. But it’s actually a Trojan horse. While the world applauds them for “protecting” science, the declaration does the opposite: it paves the way for AI labs to privatize the scientific commons.
“The Leiden Declaration doesn’t protect science. It protects the companies that want to own it.”
Let me explain. The declaration focuses on ethical principles—transparency, accountability, fairness. All good. But it deliberately sidesteps the two things that actually matter: intellectual property and data ownership. By leaving those untouched, the declaration hands AI labs a free pass. They can scrape every published paper, every raw dataset, every failed experiment, train their models on it, and then patent the resulting “discoveries.” The public pays for the research. The AI lab owns the output.
I saw this firsthand. I visited a top university where a team of physicists had spent five years collecting data on a rare quantum phenomenon. A well-funded AI company offered to “help” analyze it—free of charge, under a cozy “research agreement.” The team signed. Within months, the company had filed three patents. The physicists got co-authorship on a paper. They celebrated. They didn’t realize they’d just handed over the keys to their entire field.
“Ethics without ownership is just a permission slip for exploitation.”
This is the horror the Leiden Declaration ignores. It asks the right questions: Is the AI fair? Is it transparent? But it never asks the essential one: Who owns the future of science? By staying silent, the declaration implicitly says: the tech giants do. And that’s not a bug—it’s a feature. Many of the declaration’s signatories have ties to those same companies. They’re not naive. They’re complicit.
You might think: “But don’t we need ethical guidelines?” Yes. But not at the expense of the very structure that makes science a public good. The Leiden Declaration is a distraction. It makes us feel good while the real heist happens. The AI labs don’t care about ethics. They care about data. And the declaration just gave them a way to get it without a fight.
“If you control the data, you control the discovery. The declaration lets the data slip away.”
Here’s the twist everyone misses: the real risk isn’t that AI will make science too easy. It’s that AI will make science owned. The human struggle for knowledge—the paradigm shifts, the Eureka moments, the resistance—is being replaced by automated pattern-matching. And the declaration doesn’t just fail to stop it. It normalizes it.
I’m not saying we should ban AI in science. That’s impossible. But we need to demand that any declaration about AI and science starts with one simple rule: Any discovery made using publicly funded research data must remain in the public domain. No patents. No exclusive licenses. No ownership by algorithms.
Until that happens, the Leiden Declaration is a masterclass in misdirection. It’s a velvet glove over an iron fist. And the fist belongs to Silicon Valley.
“The next Copernicus won’t be a person. It will be a prompt. The question is: who holds the rights to the new universe?”
If we don’t rewrite the rules now, the next paradigm shift in physics won’t come from a human mind. It will come from a server farm, and we’ll only get to read the press release. That’s not science. That’s a subscription service.
FAQ
Q: Isn't the Leiden Declaration a step in the right direction for AI ethics in science?
A: It's a step, but in the wrong direction. It focuses on abstract principles like transparency while ignoring the concrete issue of who owns the data and discoveries. That's like putting a seatbelt on a car that's about to be stolen. It makes you feel safe, but you're still losing the car.
Q: What practical change would you suggest instead of the Leiden Declaration?
A: Start with a single rule: any discovery made using publicly funded research data must remain in the public domain. No patents, no exclusive licenses, no AI company owning the output. That's the only way to ensure AI accelerates science without privatizing it.
Q: Aren't you being too harsh? The declaration includes many respected academics who genuinely want to protect science.
A: Intention doesn't equal outcome. Many of those academics are unwittingly serving the interests of tech companies they have ties to. The road to a privatized scientific commons is paved with good intentions. The declaration is a classic case of the fox guarding the henhouse—except the fox wrote the ethics policy.