Artificial intelligence is no longer a future promise in medicine. At the Asia-Pacific Healthcare Quality Forum, executives from top hospitals and patient advocates from the United States, Europe, and Taiwan exchanged perspectives on what AI is already delivering inside hospitals, and what still stands between successful pilots and routine clinical use.
Patient Care with AI: Smart Lighting and Non-Contact Monitoring Improve Healthcare Services
In northern Taiwan, Liang-Kung Chen, Superintendent of Taipei Municipal Guandu Hospital, described his community hospital as a “lighthouse" for the locals. The hospital uses AI tools to coordinate services across wards, homes, and community settings, extending care beyond traditional clinical boundaries.
Take smart lighting in the hospital as an example. Chen said the system, installed in mental health wards, automatically adjusts lighting throughout the day and has improved residents’ sleep quality by 52 percent.
To improve patient safety, the hospital has also introduced smart mattresses paired with a monitoring app. Bedside falls and pressure injuries have been greatly reduced as a result, Chen said.
At larger centers, such as Taipei Veterans General Hospital (TVGH), AI has been deployed for non-contact monitoring. The hospital has adopted FaceHeart, an AI-assisted system that tracks patients’ physiology and mood using facial recognition technology, according to Wui-Chiang Lee, TVGH’s Deputy Superintendent.“ The system has been applied in both inpatient and home-care settings,” he added.
TVGH is leveraging AI across internal medicine, including in heart failure, pulmonary hypertension, hepatoma diagnosis, endoscopic image analysis, and renal anemia prediction and treatment.
Jennifer L. Bright, President and CEO of the International Consortium for Health Outcomes Measurement (ICHOM), emphasized the importance of focusing on meaningful outcomes when designing AI tools.
“While platforms can now track patient-reported data in real time, success depends on choosing the right metrics,” she said. Bright added that involving patients in designing digital systems improves engagement and builds trust, especially when data is shared back with them.
On the other hand, Oscar Gaspar, president of the European Union of Private Hospitals, cited the European Health Data Space Regulation, emphasizing the importance of giving patients access to their own health records.
C. Jason Wang, Director of the Center for Policy, Outcomes and Prevention (CPOP) at Stanford University, noted that in the United States, patients’ insurance coverage must be taken into account when considering the use of smart technologies in medical care.
Clinical Use of AI: Faster Genomic Analysis and Diagnosis Enhance Treatment Efficiency
Recognizing their core strengths and limitations, hospitals are adopting different strategies to integrate AI into daily operations and clinical care.
TVGH has pursued strategic partnerships with tech companies to access advanced tools as needed. For example, the hospital collaborated with NVIDIA to speed up genomic secondary analysis using Clara Parabricks pipelines. A task that once took 32 hours to complete now takes about 1.2 hours, a 96 percent reduction in time.
In response to the workload of clinical documentation, TVGH partnered with ASUS to develop a generative AI documentation tool. The system reduced the time required to complete inpatient medical records from approximately 7.5 minutes to just 25 seconds. Based on the hospital’s daily workload, this translates into workload savings equivalent to 4.5 full-time staff, allowing healthcare teams to devote more time to patient care and clinical decision-making.
TVGH also works with Foxconn and Microsoft on projects in telemedicine, robotics, and automation.
China Medical University Hospital has taken a different approach. According to Shih-Sheng Chang, Director of the hospital’s AI & Robotics Innovation Center, the hospital has about 50 in-house engineers working with doctors.
Chang noted that in-house engineers observe challenges in real time and design tools that meet clinicians' immediate needs. “They listen to the users: what they need and what they want to solve,” he said. This model has helped develop an electrocardiogram algorithm that shortens heart-attack treatment time by 16 minutes.
Use AI to Solve the Right Problems
Across hospitals and regions, experts at the forum agreed that technology must follow patient needs. While AI is proving its value by shortening treatment times, easing documentation workload, and helping clinicians act faster, patient welfare remains center stage.
Successful AI in healthcare starts with understanding what matters most to those receiving care. Systems built around those outcomes, rather than data collection for its own sake, can earn trust and deliver results. In the end, it’s not about having the smartest algorithm, but about solving the right problem.
