The Intersection of Technology and Humanity: Exploring the Ethics and Potential of Artificial Intelligence in Medicine

Past developments in medicine have been remarkable, which have contributed to the average life expectancy in many countries exceeding 80 years, and is expected to soon reach 100 years. However, even with advancements in medical technology, misdiagnosis can occur when patients present with nonspecific symptoms and signs. Many people around the world also do not receive appropriate healthcare due to a lack of development in its infrastructure. The integration of intelligent technology into healthcare, which can be called digital healthcare, has brought significant advancements in medical technology, such as clinical disease-supporting systems (CDSSs). Using CDSSs can provide appropriate diagnosis and treatment strategies to aid medical professionals even when the problems are not within the scope of their specialties.¹ The emergence of ChatGPT, among other innovations, is leading to further changes in the healthcare environment, including the application of artificial intelligence (AI) in diagnosis and treatment. These advancements in technology offer the possibility of early diagnosis and treatment, which would lead to improved patient outcomes.

The article published in this issue titled “Long Short-Term Memory-Based Deep Learning Models for Screening Parkinson’s Disease Using Sequential Diagnostic Codes” presents an innovative approach for screening Parkinson’s disease using deep-learning models.² The study analyzed data from the National Health Insurance Service—National Sample Cohort (NHIS-NSC) of the Republic of Korea for the time period of 2002–2013. It is a population-based sample cohort that contains information on 1,025,340 study participants, equivalent to 2.2% of the total Republic of Korea population, and the sample was randomly selected from the National Health Insurance database. The data include information on demographics, medical history, medication, medical claims, and regular health checkups. The study employed a long short-term memory (LSTM) neural network model, which was trained on sequential diagnostic codes from electronic health records to predict Parkinson’s disease. The authors demonstrated that their model outperforms traditional machine-learning methods in terms of accuracy, sensitivity, and specificity. The LSTM model achieved an accuracy of 94.25%, while the other models also showed performance with about 92% accuracy. The study has significant implications for the early detection and screening of Parkinson’s disease, which is currently challenging due to the lack of effective biomarkers. Overall, the article represents a valuable contribution to the growing body of literature on the application of deep learning models in medical diagnosis and screening. The study highlighted the potential of using electronic health records and deep learning models to improve the accuracy and efficiency of disease diagnoses.

Kim et al.³ have recently provided insight into how we can apply the technology to assess or diagnose rare or complex diseases. The authors tried to develop a screening algorithm for diagnosing orthostatic hypotension, which was popular but did not receive much attention, which resulted in underdiagnosis in medicine. Several machine-learning techniques such as support vector machines, k-nearest neighbors, and random forest classifiers were used to develop the screening algorithm that can be applied before performing the head-up tilt test on elderly patients who are unable to actively stand up by themselves and have contraindications to performing the test.

A high-quality big-data repository is essential to develop a reliable CDSS. There are still many considerations for the applicability of numerous CDSSs despite their recent development. Privacy concerns regarding data usage and judgment errors due to training datasets of inadequate quality are problems that need to be overcome gradually through social consent and technological development. Furthermore, AI replacing physicians as a profession is a current concern. However, if we agree that medicine must be among the universally applicable tremendous technological advancements that humankind has achieved, we must strive to accelerate the significant changes currently occurring in the medical environment. We hope to see more-promising research results that improve the accuracy and extend the applicability of CDSSs and AI in the field of medicine in the near future.

Availability of Data and Material

Data sharing not applicable to this article as no datasets were generated or analyzed during the study.