Revisão de escopo sobre o uso da inteligência artificial generativa no desenvolvimento do raciocínio clínico e a confecção de guia aplicado ao ensino médico

• Ano de defesa: 2025
• Autor(a) principal: Ferreira, Guilherme Freitas Bernardo
• Orientador(a): Turci, Maria Aparecida
• Banca de defesa: Toledo Júnior, Antonio Carlos de Castro, Ribeiro, Ligia Maria Cayres
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade José do Rosário Vellano
• Programa de Pós-Graduação: Programa de Mestrado em Ensino em Saúde
• Departamento: Pós-Graduação
• País: Brasil
• Palavras-chave em Português: Educação médica - racioncínio; Inteligência artificial - aplicações médicas
• Área do conhecimento CNPq: CIENCIAS DA SAUDE::MEDICINA
• Link de acesso: http://tede2.unifenas.br:8080/jspui/handle/jspui/367

The study of clinical reasoning, understood as a fundamental cognitive process in medical practice, has been the subject of research since the 1970s. The development of this essential competence requires knowledge acquisition, pattern recognition, and the development of mental scripts, making its teaching a significant challenge. With advances in artificial intelligence (AI), particularly chatbots based on large language models (LLM) such as ChatGPT, new perspectives are emerging to assist in teaching clinical reasoning. This dissertation aimed to identify how chatbots can be used to facilitate the teaching of clinical reasoning. To this end, a systematic scoping review was conducted, in which 21 publications exploring the use of these tools in medical education were selected and analyzed. The results suggest that chatbots can support learning by generating differential diagnosis lists, providing detailed explanations of clinical cases, and acting as virtual tutors in teaching simulations. Based on these findings, AI utilization possibilities were suggested grounded in educational principles that have previously been demonstrated to be effective for teaching clinical reasoning. Challenges were also identified, such as the risk of bias, the presence of inaccurate information, and the need for faculty supervision in using these tools. Building upon these results, a practical guide titled “6 Tips for Using ChatGPT to Facilitate the Teaching of Clinical Reasoning” was developed, aimed at educators. This material presents recommendations based on the reviewed literature, including suggestions for optimizing the use of ChatGPT in pedagogical activities. The guide also includes a glossary of key AI concepts, making it an accessible tool for educators who wish to integrate this technology into medical education in a critical and effective manner. Thus, this study reinforces the potential of AI as an ally in medical education, provided that it is used with pedagogical planning and attention to the inherent challenges of the technology.