OpenAI has released a policy paper advocating for AI in the life sciences, but as of yet, no AI-discovered drug has passed Phase III clinical trials

According to DynoMatics Beating monitoring, the OpenAI Policy, Research, and Science team released a report advocating for the expansion of AI in the life sciences field and shared it exclusively with Axios. The report calls for three things: opening up more access to medical and scientific data, considering advanced AI as a "national research asset," and increasing investment in "physical infrastructure" such as computing power, laboratories, and energy.

The report states that AI can accelerate drug discovery, autonomously design research tools, and compress laboratory processes from months to days. One analysis estimates that AI tools can shorten the timeline of the clinical trial phase by over 20%. OpenAI also specifically mentioned that GPT-5 Pro can identify new uses for FDA-approved drugs for diseases that currently lack effective therapies.

However, there is a clear gap between reality and vision. Currently, only a tiny number of AI-discovered or AI-designed drugs have entered the clinical trial phase, with none completing Phase III trials. A paper published in mid-2025 in Nature Medicine found that AI-discovered drugs had a failure rate in Phase II clinical trials comparable to drugs discovered through traditional methods. The authors of the paper wrote, "The question of whether AI can have a meaningful, sustained impact on drug discovery remains unanswered."

The policy demands of the report are also noteworthy. OpenAI hopes that governments will open up medical data and provide special status and resources for AI research and development at the national level, essentially lobbying for greater market access for their own products. In the U.S., it typically takes 12 to 15 years for a new drug to go from research to approval, and AI's promise to shorten this cycle is appealing to the pharmaceutical industry and regulatory agencies. However, the current clinical evidence is insufficient to support these promises. In the same week, Amazon also launched the AI drug molecule generation tool Bio Discovery, with large tech companies rushing into this field.