Work using artificial intelligence to identify antibiotics has earned Regina Barzilay and James Collins the 2026 Pritzker Prize for AI in Science Research Excellence. Their research combines deep learning, bioengineering, and experiments to search for antibacterial candidates that could act against drug-resistant pathogens.
AI helps narrow the search for antibiotics
Barzilay, MIT’s School of Engineering Distinguished Professor of AI and Health, and Collins, MIT’s Termeer Professor of Medical Engineering and Science, are faculty co-leads of the Abdul Latif Jameel Clinic for Machine Learning in Health. They published their first breakthrough paper together in 2020 with Jonathan Stokes, then an MIT postdoc and the paper’s first author.
The paper described halicin, identified by a deep learning model trained on 2,500 molecules. Halicin had first been studied as a potential diabetes drug. Researchers tested it and found that it could kill antibiotic-resistant bacteria, including Clostridium difficile, Acinetobacter baumannii, and Mycobacterium tuberculosis.
The discovery process took a few weeks, where it had once taken years. The result showed how AI could help researchers identify promising candidates more quickly amid concern about antimicrobial resistance. It did not remove the need for testing: the work paired computational screening with experimental validation.
Two discoveries show different approaches
After halicin, Barzilay and Collins worked with Stokes, now an assistant professor at McMaster University, on abaucin, identified in 2023. It is a narrow-spectrum antibiotic that demonstrated effectiveness against A. baumannii, a treatment-resistant bacterium.
Abaucin’s narrower reach is presented as an advantage. Targeting a single bacterial species may reduce the likelihood of resistance developing and limit harm to beneficial bacteria in the gut. The contrast with halicin illustrates that antibiotic discovery can pursue candidates with different scopes, depending on the pathogen researchers aim to address.
The award recognizes a cross-disciplinary effort
The Pritzker Prize for AI in Science Research Excellence was established in 2024 by the University of Chicago in partnership with Caltech. It recognizes contributions that jointly advance AI and the natural sciences or engineering. The prize citation recognizes Barzilay and Collins for “pioneering AI-enabled antibiotic discovery, integrating deep learning, bioengineering, and experimental validation to identify novel antibacterial candidates against drug-resistant pathogens.”
That combination matters because identifying a candidate is only one part of the work. Machine learning can help search through molecules, while bioengineering and experiments help determine whether a candidate has antibacterial activity. The prize highlights the collaboration across these disciplines and the researchers who brought them together.
AI's role has limits
Collins described a shift in how AI is used, from a discovery aid toward a design tool. “Historically, antibiotic discovery has depended on laborious, trial-and-error strategies,” he said. He added that generative AI could open new directions for antibiotic discovery.
Barzilay emphasized that the broader process remains at an early stage. “With AI, we can only automate parts of the pipeline, rather than totally transforming the discovery process.” She said progress will require realistic AI models of human biology. The distinction is important: faster candidate identification can help, but the source does not suggest that AI alone completes drug discovery.
Halicin and abaucin offer examples of what machine learning can contribute: finding candidates with activity against resistant bacteria, including a focused candidate against A. baumannii. The award recognizes that work while the researchers’ own comments point to the remaining challenge of building models that better represent human biology.