Scientists at the U.S. National Library of Medicine developed GeneAgent, an artificial intelligence tool leveraging large language models (LLMs) with autonomous self-verification against expert-curated biological databases. Published in Nature Methods, GeneAgent improves the accuracy of gene-set functional analyses by mitigating hallucinations and circular reasoning inherent in LLMs. This advancement enhances computational genomics workflows by providing verified biological insights, supporting more reliable interpretations of high-throughput gene data in research and drug development.