چالش های استفاده از مدل های زبانی بزرگ در مراقبت های بهداشتی
کد: G-1858
نویسندگان: Zahra Emami ℗, Alireza Sohrabi, Leila Shahmoradi *
زمان بندی: زمان بندی نشده!
برچسب: دستیار مجازی هوشمند
دانلود: دانلود پوستر
خلاصه مقاله:
خلاصه مقاله
Introduction: Large Language Models (LLMs) play a significant role in generating AI-driven content. However, without proper oversight and responsible design, they can become tools for disseminating misinformation. This study examines the challenges of LLMs in healthcare and proposes strategies for their responsible use. Materials and Methods: This review study was done in 2025, by searching databases such as ScienceDirect, PubMed, and Google Scholar with the keywords “LLMs”, “Ethics”, “Healthcare” and “ChatGPT” with a time limit of three years. Sources were reviewed to ensure comprehensive search results, the title and abstract of the articles were reviewed, and the articles that met the inclusion criteria were included in the study. Results: Review of Studies shows that AI challenges in healthcare can be categorized into ethical, technical, and social area. Ethical concerns include accountability, ambiguity in responsibility allocation, preserving human dignity, bias, privacy protection, intellectual property rights, and issues like generating texts without proper citations or incorrect references. Technical limitations involve dependence on training data, lack of linguistic understanding, inability to automatically evaluate generated content, producing convincing yet incorrect responses (hallucinations), and reliance on probabilistic algorithms. Social concerns encompass reduced human oversight, altered physician-patient relationships, and diminished transparency. To address these challenges, proposed solutions include implementing response validation processes, leveraging blockchain technology, ensuring data privacy and security, obtaining informed consent, using more diverse and representative datasets, fostering interdisciplinary research, maintaining continuous human oversight, integrating AI-related topics into educational programs, and enhancing user experience through personalization and feedback mechanisms. Conclusion: The use of large language models in healthcare can improve service delivery and enhance the quality of information. However, their challenges must be carefully managed. Implementing the strategies suggested in this study will help optimize LLM performance and ensure compliance with ethical principles.
کلمات کلیدی
Large Language Models, Ethics, ChatGPT, Healthcare
