부산대학교 도서관

This article explores the current state of artificial intelligence (AI) in the medical field, specifically in oncology. It acknowledges the hype surrounding AI but also highlights its limited presence in clinical practice. The article compares predictive modeling using AI to traditional evidence-based medicine scoring algorithms and emphasizes the importance of domain knowledge and expert opinion in developing these models. It also discusses challenges such as reproducibility and the lack of clear standards for evaluating AI medical models. The article concludes by discussing the potential requirements for AI models to reach the level of relevance seen in traditional scoring systems. While AI and LLMs have the potential to improve patient care, their ability to do so efficiently and with fewer mistakes is uncertain, and further research and evaluation are needed. [Extracted from the article]