One of the most touted promises of medical artificial intelligence (AI) is that they can help human clinical doctors interpret images such as X-rays and CT scans more accurately, making more accurate diagnostic reports and enhancing the performance of radiologists. But is this actually the case? A collaborative study between Harvard Medical School, Massachusetts Institute of Technology, and Stanford University in the United States has shown that the effectiveness of using AI tools for image interpretation seems to vary among clinical physicians. In other words, whether it is beneficial or useless is still up to mankind at this stage. Because research results indicate that differences in individual clinical physicians can affect the interaction between humans and machines in key ways that AI experts have not fully understood. This analysis was recently published in the journal Nature Medicine. Considering the personal factors of doctors, research has shown that in some cases, the use of AI may interfere with the performance of radiologists and affect the accuracy of their explanations. Although previous studies have shown that AI assistants can indeed improve the diagnostic performance of doctors, these studies view doctors as a whole and do not consider the differences between different doctors. In clinical practice, every doctor's judgment is 100% accurate for patients. In contrast, this new study focuses on personal factors of clinical doctors - professional field, years of practice, and previous experience using AI tools - and analyzes how these factors play a role in human-computer collaboration. Researchers analyzed how AI affects the performance of 140 radiologists in 15 X-ray diagnostic tasks, where doctors need to reliably identify prominent features on images and make accurate diagnoses. This analysis involves 324 patients with 15 different conditions. In order to determine how AI affects doctors' ability to detect and correctly identify problems, researchers used advanced computational methods to obtain changes in performance when using and not using AI. The results showed that the effectiveness of AI assistance varied among radiologists, with some radiologists showing improved performance due to AI while others showed "deterioration". Assistant Professor of Biomedical Informatics at the Brawanik Institute of the Royal College of Medicine, Paranavus Lapcole, confirmed the research team's findings and stated that "we should not view doctors as a unified group and only consider the 'average' impact of AI on their performance.". However, this discovery does not mean that doctors and clinics should be prevented from adopting AI. On the contrary, the results indicate the need for a better understanding of how humans and AI interact and the design of carefully calibrated methods to improve rather than harm human performance. The AI "assistant" is still difficult to predict. Given that imaging is considered the clinical medical field that can receive the greatest assistance from AI, the results of this study are quite representative. It is worth noting from this discovery that in radiology, AI is surprisingly influencing the performance of human doctors. For example, contrary to the expectations of researchers, factors such as the years of experience of radiologists, whether they specialize in chest radiology, and whether they have previously used AI devices cannot reliably predict the impact of AI tools on their work performance. Another challenge is the universal view