Federal researchers use LLMs as medical research copilots

Federal medical researchers are turning to large language models to manage the growing volume of scientific information. Cancer-related publications in the National Library of Medicine’s PubMed database have more than doubled since 2005, prompting the National Cancer Institute to build NanCI, a mobile and web application launched in 2024 that summarizes findings, recommends papers and helps researchers identify potential collaborators.

NCI later added Ask NanCI, a conversational feature backed by Google Cloud infrastructure and LLM access, allowing researchers to query papers and figures directly. The tool reflects a broader shift toward AI systems that support discovery, collaboration and research direction rather than simply storing scientific information.

Other agencies are emphasizing secure deployment and data governance. Argonne National Laboratory launched Argo in 2024 to provide protected access to models including Google’s Gemini and OpenAI’s GPT-5.4, while keeping communications on the lab’s network. The FDA built its Elsa app with an application layer between users and models, paired with work to consolidate 40 application and submission systems into one platform. Currently, 80% of FDA employees use Elsa for tasks such as reviewing direct-to-consumer ad submissions, with experts retaining final judgment.