Case Study: NIH Cuts Literature Review Time with AI from ICF

NIH cuts literature review time 6x with ICF

The National Institutes of Health (NIH) faced a slow and resource-intensive challenge in manually reviewing thousands of scientific papers for its Complement-ARIE program. To accelerate this process for literature on new approach methods (NAMs), NIH partnered with vendor ICF for a faster, AI-driven solution.

ICF implemented a reproducible and scalable GenAI tool that integrated its Litstream product and Anthropic’s Claude V2 model. This solution from ICF achieved 91% accuracy and reduced the literature review timeline from six months to just six weeks, cutting required labor by 95% and earning NIH’s Director’s Award.

National Institutes of Health (NIH)