AI, biomedicine and the NIH

The relevance and potential of artificial intelligence (AI) for biology and medicine should not be judged by NIH attitudes toward the field. However, NIH is the principal funding source for biomedically relevant R&D, and its enthusiasm (or apathy) for a domain can significantly alter the pace of development and the projects to which scientists devote their time. The notorious "AI winter", usually attributed to the disappointment of exaggerated expectations, more or less froze funding for AI development in general. Biomedicine was not spared, and the relatively small but enthusiastic funding provided by the National Library of Medicine dried up in the 1990s. NIH as a whole has never embraced AI, and in fact, did not develop much enthusiasm for computation in general until 2000. Even today, enthusiasm for AI (broadly defined) is modest. However, the emergence of new and complex problems for clinicians and biologists, as well as the federal interest in electronic health records, has stimulated some renewed interest in the potential of AI. Some AI can be seen in a number of funded projects, including the Strategic Health IT Advanced Research Projects (SHARP) Program, the Clinical and Translational Science Awards, and the National Centers for Biomedical Computing. The National Library of Medicine recently awarded 13 research contracts for projects in "computational thinking" relevant to biomedicine. Funding opportunities for AI projects are becoming available at NIH for applicants who can identify important problems, know where to look for support, and wordsmith with sufficient shrewdness.