Charlottesville Profs and Pints

Profs and Pints Charlottesville presents: “When AI Meets Biology,” on the benefits and risks of an artificial intelligence-driven revolution in biological research, with Stephen Turner, associate professor and assistant dean for research in the University of Virginia’s School of Data Science and geneticist with extensive experience with biotech firms.

[All tickets must be purchased online, with sales tax and processing fees added, at https://events.ticketleap.com/tickets/profsandpints/charlottesville-ai-biology .]

In February the AI company OpenAI and the biotech company Gingko Bioworks announced the results of 35,000 biological experiments designed and run by a relatively new player in the biotech field: the artificial intelligence model GPT 5. Based on goals that humans had set, the AI model designed experiments that it had carried out mainly by robots, with the resulting data fed into the AI model to inform new rounds of experimentation.

The approach produced a desired protein at exceptionally low cost. It also raised a troubling question: Are we involving AI in biological experiments to an extent that we, as humans, might come to regret?

Hear these and other questions related to AI’s involvement in biological research tackled by Dr. Stephen Turner, an expert on AI and biosecurity and advisor to biotech startups, including the one that recently sought to bring the dire wolf back from extinction. He brings to the talk extensive experience at the intersection of academia, industry, government, policy, and entrepreneurship.

Dr. Turner will start by looking at the remarkable achievements of AI in biology, describing how AI systems can now design experiments, order lab supplies online, and ace college-level biology exams, with some AI models scoring higher than human experts on virology and biosecurity tests.

He’ll then look at what happens when the rubber meets the road and AI gets involved in “bench research”—the actual work of carrying out work in the lab, where skills like pipetting technique and gel reading matter more than the ability to recite facts. You’ll learn about research showing that frontier AI failed to be meaningfully better than humans at such tasks.

That gap—between what AI appears to know and what it can actually do in the real world--is at the center of a growing policy crisis. Policymakers are already making decisions about which AI models to restrict, how to screen DNA synthesis orders, and what counts as “safe enough,” with all such decisions being based on benchmark scores that no one has validated against real laboratory outcomes.

Meanwhile, a newer threat is emerging: Cloud laboratories and autonomous lab platforms could enable someone to skip bench research entirely, by outsourcing the physical work to robots or third-party services. The old assumption that “hands-on skill” and human involvement protects us starts to erode if AI can write the research protocol and entrust a third party to run it.

You’ll emerge from the talk with an appreciation of the need for human involvement in such research discussions. (Advance tickets: $13.50. Doors: $17, or $15 with a student ID. Doors open for talk attendees at 4:30 pm and the talk starts at 6 pm.)
Image by Canva.