Elsevier Introduces Evaluation Framework for AI in Clinical Tools

Elsevier has introduced a new evaluation framework for assessing the performance and safety of generative AI-powered clinical reference tools, announced in a press release. This framework is designed to evaluate AI-generated responses across five critical dimensions: query comprehension, response helpfulness, correctness, completeness, and potential for clinical harm.

The framework, which will be featured in an upcoming issue of the Journal of the American Medical Informatics Association (JAMIA), aims to ensure that AI tools provide accurate and relevant information to healthcare professionals. Initial evaluations of Elsevier's ClinicalKey AI showed high accuracy, with 94.4% of responses rated as helpful and 95.5% assessed as completely correct.

Elsevier's initiative represents a significant step in the responsible integration of AI technologies in healthcare, setting a new standard for AI deployment in clinical settings. The company is also actively contributing to industry-wide standards for responsible AI use through its partnership with the Coalition for Health AI.