Exclusive: OpenAI lobbies for expanded AI role in life sciences
Add Axios as your preferred source to
see more of our stories on Google.
Illustration: Gabriella Turrisi/Axios
Advances in AI's ability to take on novel scientific work are helping researchers move faster, connect siloed knowledge, and design treatments more efficiently, according to a new report from OpenAI's policy, research and sciences team shared first with Axios.
Why it matters: The life sciences have saved hundreds of millions of lives over the past century, but progress has slowed dramatically — even as the toughest diseases remain unsolved.
The big picture: Biomedical discovery moves at a snail's pace.
- In the U.S., bringing a new drug from research to approval often takes 12 to 15 years.
- "Eroom's Law" — Moore's Law in reverse — holds that scientific research slows as knowledge accumulates, because each new generation of scientists must spend more time just absorbing what came before.
- AI has nothing but time.
Zoom in: AI can speed up drug discovery, design new tools itself and automate labs to compress experiments from months to days, OpenAI says.
- One analysis estimated that these tools could cut clinical-phase timelines by more than 20%.
- GPT-5 Pro, OpenAI argues, can find new uses for FDA-approved drugs in diseases that still lack solid treatments.
State of play: Competition with China and policy shifts in the U.S. are raising pressure to speed drug development — making AI appealing to big pharma, Axios Pro biotech reporter Katherine Davis wrote last year.
- This week, Amazon launched Bio Discovery, an AI agent that generates and evaluates potential drug molecules.
Reality check: OpenAI's report functions as a policy pitch, arguing for changes that would support broader AI use in life sciences.
- It calls for greater access to medical and scientific data, treating advanced AI as a national research resource, and pushes for investment in the "physical stack," meaning compute, labs, energy and infrastructure.
Yes, but: Only a few AI-discovered or AI-designed drugs have reached clinical trials.
- AI-discovered drugs have experienced similar levels of phase 2 trial failure as non-AI-discovered drugs, according to a mid-2025 Nature Medicine paper.
- No fully AI-discovered or AI-designed drug has made it through phase 3 trials.
- "The question of whether AI can impart meaningful, sustained disruption to drug development has remained unanswered," the researchers of the Nature paper write.
AI hallucinations, bias and other inaccuracies are not as common as they once were, but the models are far from error-free.
- OpenAI says they "rigorously" measure accuracy across scientific domains.
- The company says it views AI as a tool to help scientists synthesize evidence, generate hypotheses, and support analysis, but not as a replacement for expert judgment or real-world validation.
The bottom line: AI could speed up slow, expensive, failure-prone drug development — but the proof is still in the pipeline.
