From black box to clear box: explainable AI for next-gen pharmacovigilance.

Key Points

• The research aims to explore the integration of explainable AI in pharmacovigilance to improve transparency and safety signal detection.
• Review of EU AI Act and its implications for pharmacovigilance (PV) algorithms.
• Analysis of the FDA's Good Machine Learning Practice (GMLP) principles.
• Assessment of India's Responsible AI principles and their relevance to PV.
• Recommendation for a regulatory-fit explainable AI (XAI) framework.
• Identified the need for increased human oversight in AI-driven pharmacovigilance.
• Highlighted potential for improved safety signal detection through explainable AI methods.
• Proposed convergence of governance and human oversight with XAI for enhanced international trust.