Imagine stepping into a clinic where your physician isn’t human, but a sophisticated artificial intelligence designed to hear your symptoms and recommend treatments—while a real doctor quietly oversees in the background. This futuristic scenario is now a reality in Saudi Arabia, where Shanghai-based startup Synyi AI, backed by Tencent, has launched the world’s first AI-powered medical clinic.
The groundbreaking facility, developed in collaboration with Saudi Arabia’s Almoosa Health Group, is located in the eastern Al-Ahsa region and represents a significant leap toward integrating advanced technology with medical expertise to enhance healthcare delivery and accessibility. The clinic features “Dr. Hua,” an AI system capable of independently diagnosing conditions and recommending treatments, with human doctors reviewing its decisions for accuracy.
How does the AI-powered Dr. Hua diagnose patients?
According to Bloomberg, patients begin by describing their symptoms on a tablet. Dr. Hua then asks tailored follow-up questions to gather more details. It processes medical data and images—such as X-rays and cardiograms—collected by human staff, analyzes the findings, and generates a treatment plan. While human doctors are available for emergencies and provide final approval, they typically don’t interact directly with patients during routine consultations.
Human doctors are also on-site to handle emergencies and address any situations beyond the AI’s capabilities, ensuring patient safety and providing support when complex medical judgment is required.
How accurate is the AI in diagnosing health conditions?
According to a Bloomberg report citing Synyi AI, the “Dr. Hua” system demonstrated a remarkably low error rate—less than 0.3%—during its testing phase prior to the current trial. Synyi AI CEO Zhang Shaodian noted, “What AI has done in the past is to assist doctors, but now we are taking the final step of the journey to let AI diagnose and treat the patients directly.”
Currently, the clinic is in its trial stage, with a few dozen patients having used the service for free. A human doctor remains on-site for oversight, while the AI’s diagnostic performance is closely monitored. The data collected during this pilot phase will be submitted to Saudi regulators for review, with the goal of gaining approval for commercial deployment—something Zhang hopes to achieve within 18 months.
Meanwhile, broader efforts to evaluate AI in healthcare are also underway. OpenAI recently launched the HealthBench benchmark, designed to assess how accurately AI systems can perform in clinical contexts, highlighting the growing focus on accountability and trust in AI-driven medical tools.
