ChatGPT的放射肿瘤学委员会考试。
A Radiation Oncology Board Exam of ChatGPT.
作者信息
Barbour Andrew B, Barbour T Aleksandr
机构信息
Radiation Oncology, University of Washington - Fred Hutchinson Cancer Center, Seattle, USA.
Starlink, SpaceX, Redmond, USA.
出版信息
Cureus. 2023 Sep 1;15(9):e44541. doi: 10.7759/cureus.44541. eCollection 2023 Sep.
As artificial intelligence (AI) models improve and become widely integrated into healthcare systems, healthcare providers must understand the strengths and limitations of AI tools to realize the full spectrum of potential patient-care benefits. However, most providers have a poor understanding of AI, leading to distrust and poor adoption of this emerging technology. To bridge this divide, this editorial presents a novel view of ChatGPT's current capabilities in the medical field of radiation oncology. By replicating the format of the oral qualification exam required for radiation oncology board certification, we demonstrate ChatGPT's ability to analyze a commonly encountered patient case, make diagnostic decisions, and integrate information to generate treatment recommendations. Through this simulation, we highlight ChatGPT's strengths and limitations in replicating human decision-making in clinical radiation oncology, while providing an accessible resource to educate radiation oncologists on the capabilities of AI chatbots.
随着人工智能（AI）模型的改进并广泛融入医疗保健系统，医疗保健提供者必须了解AI工具的优势和局限性，以实现潜在的患者护理益处的全范围。然而，大多数提供者对AI了解甚少，导致对这项新兴技术的不信任和采用率低。为了弥合这一差距，本社论提出了对ChatGPT在放射肿瘤学医学领域当前能力的新观点。通过复制放射肿瘤学委员会认证所需的口头资格考试的形式，我们展示了ChatGPT分析常见患者病例、做出诊断决策以及整合信息以生成治疗建议的能力。通过这种模拟，我们突出了ChatGPT在临床放射肿瘤学中复制人类决策方面的优势和局限性，同时提供了一个易于获取的资源，以教育放射肿瘤学家关于AI聊天机器人的能力。
Zotero 插件