SNOMED International's position statement makes the case that AI in healthcare is only safe when its outputs are constrained and governed by verified clinical terminology — and outlines guiding principles and internal AI initiatives through which SNOMED CT serves as that essential guardrail.

SNOMED International, in collaboration with Data Driven, has developed an Entity Linking Challenge and Benchmark to advance the automatic mapping of unstructured clinical text to verified SNOMED CT concepts — setting a shared, validated standard against which AI and NLP systems can be rigorously tested and compared.

Learn about the AI Benchmark

The goal of developing AI-enabled translation tools is to accelerate the localisation of SNOMED CT into low-resource languages, allowing member countries to produce accurate, concept-anchored translations at scale — without compromising the clinical precision that safe implementation demands.

As SNOMED CT grows in size and complexity, AI-assisted quality assurance offers a powerful way to help terminologists manage increasing workloads — automating the detection of issues, generating tasks directly, and ultimately supporting the concept modelling process itself.