03/06/2026
Delighted to share our project on AI assurance testing for a clinical guidance chatbot at IMDA’s Sandbox booth and Sandbox Showcase during the Asia Tech x Singapore Conference 2026 (ATxSG) last month.
The chatbot allows doctors to query clinical guidelines on diagnosis, treatment protocols, medication dosing, and patient care management. With possible implications on clinical decision-making, testing focused on ensuring that the chatbot remained accurate, safe, reliable, and resilient under realistic clinical conditions. A multi-disciplinary team including senior doctors, senior residents, and pharmacists with different domain knowledge played a critical role in validating outputs, refining responses, and ensuring the technology aligns with clinical practice and patient safety standards.
CGH Chief Data and Digital Officer and Head of Cardiology, Clin Assoc Prof Chow Weien, Lead Principal Analyst Dr Srinath Sridharan, and project collaborator Guardrails AI shared insights from the testing journey with fellow practitioners during the Sandbox Showcase.
“CGH approaches AI implementation with focused, purposeful deployment and strong human oversight. This includes rigorous testing of AI models and post deployment surveillance. This is the reason why we developed the chatbot grounded in a clinical knowledge base,” said Clin Assoc Prof Chow, who is also Chair of CGH’s AI and Digital Committee and Principal Investigator of the project.
“Due to the comprehensive and rigorous testing approach, we were able to observe patterns that allowed us to further refine our clinical guidance chatbot," he added.
Proud to contribute to ongoing conversations on responsible AI innovation and AI assurance in healthcare as we continue to 👏
