Medical AI Moves Beyond Disease Detection and Diagnosis to Patient Care as the "Care AI" Era Begins

Kang Minsu, Eulji University Professor: "AI should enable medical staff to focus on patients"
Connecting everything from risk detection and prediction to intervention and post-discharge care

There have been calls for medical artificial intelligence (AI) to move beyond detecting diseases and assisting with diagnosis, expanding into "Care AI" that connects patient treatment and recovery with care after discharge. The idea is that AI can lead to tangible improvements in medical services only if it not only detects changes in patients' conditions and predicts risks, but also connects those insights to medical staff's decisions and follow-up actions.

Kang Minsoo, professor at Eulji University's College of Advanced Studies

Kang Minsu, professor at Eulji University's College of Advanced Studies, made the remarks while delivering a lecture titled "Using Artificial Intelligence in Patient Care Systems" at a medical symposium marking the 70th anniversary of Eulji University Medical Center, held on October 30 at The Grand Lotte Seoul in Jung-gu, Seoul.

"The purpose of future medical AI is not technology, but better patient care," Kang said. "While conventional medical AI has focused on what to detect, we now need to consider what to do after something is detected."

He explained that medical AI should evolve from "Diagnostic Support AI," which detects, classifies and predicts signs of disease, into "Care AI," which connects the process and outcomes of patient care. While Diagnostic Support AI answers questions at a particular point in time, Care AI alerts users to risks, connects them with necessary interventions and tracks the results.

As society ages rapidly, demand for patient care is also becoming more diverse. In addition to the risks of polypharmacy, falls and readmission among older patients, medical institutions must address challenges such as long-term recovery management for rehabilitation patients and language and cultural barriers for foreign patients.

Kang proposed using AI to improve continuity of care by reducing gaps in explanation and understanding between medical staff and patients, between changes in patients' conditions and medical staff's awareness of them, and between the detection of risks and actual intervention. For example, if an abnormality is detected in a ward patient's vital signs, AI could analyze the risk of deterioration and alert medical staff, who could then reassess the patient's condition and decide on the necessary measures.

"Hospitals already have a great deal of data," Kang said. "The important thing is not simply to display the data, but to identify and provide the information that medical staff need to focus on."

He also predicted that the scope of patient care would expand beyond hospitals as generative AI, AI agents and physical AI become integrated. Generative AI could summarize medical records and draft documents, while AI agents could connect follow-up tasks such as alerts, referrals and patient education based on medical staff's decisions. Combining these technologies with robots, smart devices and wearable devices could also create an environment in which patients' vital signs and activity levels are monitored and care is supported at home.

However, medical decisions and high-risk interventions must always be made under the review and control of medical staff, and as the scope of medical AI applications expands, he noted that ensuring safety and accountability will also become a key issue, including through safety validation, clinical evidence, personal information protection and interoperability between systems.

"AI's value lies in giving medical staff more time to spend face-to-face with patients," Kang said. "The future should not be one in which AI replaces people, but one in which it helps people care for other people better."

This content was produced with the assistance of AI translation services.