Federating AI-related regulations for human therapeutics: an AI-enabled, continuously updating regulatory...

Area of Science:

• Healthcare Regulation
• Artificial Intelligence
• Drug Development

Background:

• Global AI in healthcare regulation is fragmented, with over 1,000 policies across 70+ countries, primarily from high-income nations.
• Major regulatory bodies (FDA, EMA, WHO) have issued AI guidance, but a unified comparison of their approaches to AI-enabled therapeutics is lacking.
• This fragmentation can lead to interpretation disparities, potentially delaying worldwide access to novel AI-based therapies.

Purpose of the Study:

• To address the lack of unified global AI healthcare regulation by developing an AI-enabled, continuously updated regulatory intelligence system (AICURIS).
• To enable structured comparison and identification of alignment/divergence in AI-related regulatory content across jurisdictions.
• To facilitate evidence-informed discussions on regulatory convergence for AI-enabled therapeutics.

Main Methods:

• Developed AICURIS, an AI-enabled regulatory intelligence system trained on AI regulations from sentinel authorities (FDA, EMA, WHO).
• Utilized an AI-enabled hybrid semantic similarity and keyword scoring model on over 400,000 regulatory documents (since 2019).
• Achieved 95% recall for high-confidence AI-related content, analyzing document composition, model optimization, performance, validation, bias mitigation, and real-time monitoring.

Main Results:

• The AI-enabled hybrid model demonstrated high recall (95%) in identifying AI-related regulatory content.
• Analysis of FDA, EMA, and WHO documents revealed key findings across six parts: corpus composition, model optimization, classification performance, ensemble validation, bias mitigation, and monitoring.
• The system effectively federates AI-related drug regulations across jurisdictions.

Conclusions:

• AI-powered, data-driven approaches can effectively unify disparate AI drug regulations globally.
• AICURIS reduces global regulatory disparities, fostering more equitable and efficient innovation in human therapeutics.
• The system supports collaborative development and accelerates access to novel AI-based therapies worldwide.