Children experience pain and distress in clinical settings every day. The negative consequences of unaddressed pain can be both short-term (e.g., fear, distress, inability to perform procedures) and long-term (e.g., needle phobia, anxiety, healthcare overuse or avoidance). Socially assistive robotics (SARs) is an area of robotics where the goal is for a robot to use speech, facial expressions, and gestures to interact with a human partner for the purpose of providing assistance. The utility of SARs in pediatric settings have shown promise, demonstrating high acceptance among children and efficacy in reducing procedure-related distress. However, existing studies are all hindered by a critical technical limitation: all robots were operated remotely and employed entirely scripted behaviour with very limited real-time responsiveness and complete lack of autonomy. Artificial intelligence (AI), the ability for computer systems to make autonomous decisions and independently select appropriate behaviours, has potential to address these limitations and improve the clinical experience for children and their families.    Our research group brings together a multidisciplinary, international team of experts in healthcare, biomedical sciences, social sciences, humanities, natural sciences and engineering to develop and evaluate a clinically relevant and responsive AI-enhanced social robot. In this symposium, we will share our experiences working an a multi-disciplinary international team using a user centred co-design to create the robot system and the ethical implications of using AI in healthcare.