Post by Yale School of Engineering & Applied Science

AI tools are already reshaping how scientists work, helping researchers sift through vast bodies of literature, summarize findings, and generate hypotheses. But in high-stakes scientific settings, the limitations are real. These systems can overlook key publications, cite weak evidence, or express confidence in claims that aren't supported by current science. Yale Engineering's Arman Cohan has received an NSF CAREER Award to address that directly. His work focuses on building more trustworthy, evidence-grounded AI for scientific research – developing rigorous evaluation frameworks, better methods for adapting large language models to scientific domains, and tools that make AI outputs more transparent and interpretable. The goal, according to Cohan, is understanding when these systems can be trusted, where they fail, and how to make them more reliable across scientific disciplines. Read more: https://loom.ly/spFu5Z4 #YaleEngineering #WhatsNext #AI #NSFCAREERAward #Yale #ComputerScience #Research