AI? Let's first talk about FHIR, SNOMED CT, quality, and bias
AI? Let's first talk about FHIR, SNOMED, quality, and bias
Summary
Dr. Carina Vorisek, a digital health innovation leader, highlights 'data interoperability' as the primary obstacle to AI adoption in healthcare. She notes that LLMs struggle to structure unstructured text into standardized medical terminology, emphasizing that solving data standards and interoperability is crucial for practical AI deployment.
Details
This article, presented as an interview/Q&A with Dr. Carina Vorisek, discusses the foundational requirements for adopting Artificial Intelligence (AI) in healthcare. Dr. Vorisek points out that most current AI applications are siloed—developed and deployed within individual institutions. She stresses that interoperability is fundamental for safe and meaningful AI adoption. Specifically, she cautions that Large Language Models (LLMs) are not yet capable of addressing the challenge of structuring unstructured text into standardized medical terminology. Therefore, achieving widespread AI adoption requires solving deep-seated issues related to data standards and system integration. The discussion highlights the necessity of robust data standards (like FHIR) and systemic interoperability before advanced AI can be safely or effectively utilized in clinical settings.
Technology Note
FHIR（Fast Healthcare Interoperability Resources）は医療データ交換の国際標準。このエントリの関連技術: FHIR
Original content copyright by respective publishers