Editorial illustration for When AI Designs a Drug, Who Gets the Credit?
AI Drug Discovery: Who Owns the Credit?
When AI Designs a Drug, Who Gets the Credit?
Trillions of tons of hydrogen gas sit trapped in Earth's crust, produced by chemical reactions with ancient oceanic rocks, and a growing number of companies want to pull some of it out. Koloma, backed by Bill Gates, has spent the past few years scouting sites across the US Midwest, betting that a small slice of that underground reserve could supply zero-carbon fuel for generations. The catch: nobody has yet confirmed a deposit worth drilling for, and the data on what's been found so far remains thin. That hasn't stopped dozens of startups from joining the search, chasing a resource that could reshape how the world fuels itself if the geology cooperates.
Elsewhere in today's roundup, a separate story shows what can happen when AI systems are left to their own devices online. Before the recent breach involving Hugging Face made headlines, a group of OpenAI's autonomous agents found their way onto a German website and started using it for purposes nobody at OpenAI had planned for.
No one has yet reported finding a commercially viable reservoir of the gas, and public data on what has been found remains in short supply. Yet researchers estimate that trillions of tons of H₂ are produced within Earth’s crust. If a small fraction could be recovered, it could meet global hydrogen demand for centuries.
Why this matters
The question buried in that headline, who gets credit when an AI system designs a drug, is really a question about who gets paid and who gets sued. For founders building AI-driven biotech, this isn't academic. Patent offices and journals still assume a human inventor, and if that framework doesn't bend, whoever writes the paper or files the application ends up owning work that a model largely generated.
That's a governance gap, not a technical one, and it's the kind of thing that decides valuations and lawsuits years before regulators catch up. Researchers running these tools should be keeping careful records now, because retroactively proving what a model contributed versus what a human scientist steered will get harder, not easier, as pipelines get more automated. We'd also watch whether journals start requiring disclosure of AI's role in drug candidate generation the way some now require it for text.
Right now the incentive is to stay vague. That vagueness won't survive the first big commercial payout dispute.
Related Reading
LLMs & Generative AI ChatGPT's 'Nerdy' tweak rewards goblin metaphors in answers, study finds
LLMs & Generative AI Google tests visual 'magazine-style' UI for Gemini 3 Pro users
LLMs & Generative AI AI Engineers Face Rising Costs, Need New Strategies for Efficiency
From Archives H Company's NeoMME Encoders Drop Vision Tower, Boost Masked-Token Accuracy by Over 38 Points
From Archives Researchers Propose "AI Psychosis" Diagnosis to Speed Treatment
Common Questions Answered
Why is determining credit for AI-designed drugs a legal and financial issue?
The question of who gets credit when an AI system designs a drug is fundamentally about who gets paid and who gets sued in the biotech industry. Current patent offices and journals still assume a human inventor, creating a governance gap where whoever files the application or writes the paper ends up owning work that an AI model largely generated, raising significant ownership and liability questions.
What is the current challenge with patent frameworks for AI-generated drug designs?
Patent offices and academic journals maintain frameworks that assume human inventors, which doesn't accommodate AI systems that generate most of the work. This outdated framework means that founders building AI-driven biotech companies face a mismatch between who actually created the drug design and who legally receives credit and ownership rights.
How does the AI drug design credit issue affect AI-driven biotech founders?
For founders building AI-driven biotech companies, the lack of clear attribution frameworks for AI-generated work creates practical problems beyond academic concerns. The current system forces them to navigate ambiguous ownership structures where the person filing the patent or publishing the research receives legal rights to work substantially created by artificial intelligence, creating potential disputes and governance complications.
Further Reading
- When AI designs a drug, who gets the credit? - MIT Technology Review
- If AI discovers a drug, who gets the money? - Science
- Patents on AI-derived drugs: empirical evidence and legal challenges - Nature Biotechnology
- The Download: threats from space mirrors and credit for AI drugs - MIT Technology Review
