Medical AI and Thai Medical Device Law: Limits of the Medical Device Act B.E. 2551 in Regulating AI that Shapes Physicians' Clinical Decisions

Abstract

This article examines whether and to what extent Thai medical device law can regulate medical artificial intelligence (AI) systems that influence physicians' clinical decisions. The objectives are: (1) to examine the scope of the Medical Device Act B.E. 2551 (2008), as amended by the Medical Device Act (No. 2) B.E. 2562 (2019), in regulating medical software and AI; (2) to analyse Thai Food and Drug Administration (Thai FDA) guidance on Software as a Medical Device (SaMD), including the June 2024 issue and the October 2024 revision; (3) to evaluate the limits of the Thai framework when AI directly or significantly shapes clinical decisions; and (4) to compare the European Union's Regulation (EU) 2024/1689 and selected U.S. Food and Drug Administration guidance, in order to propose regulatory directions for Thailand. The study uses documentary legal research and functional comparative analysis.

The findings are fourfold. First, Thai law already provides a legal basis for treating some medical AI as medical devices because Section 4 includes "products, software, or any other items" and turns on the intended medical purpose. Second, Thai FDA guidance systematically addresses screening, risk classification, registration, and software modification, covering AI for diagnosis, screening, risk scoring, triage, and treatment planning. Third, existing regulations may still be insufficient when AI strongly shapes clinical judgment, particularly regarding human oversight, transparency, data quality and bias, real-world performance monitoring, model updates, and the allocation of responsibility. Fourth, comparative law confirms that effective regulation must address both product attributes and the system's effect on human decision-makers.

The article argues that Thailand should build on the medical device framework while developing high-risk medical AI guidance that treats clinical influence, human reliance, and lifecycle governance as central regulatory factors.