Comparative Evaluation of Diagnosis and Treatment Plan Given by Pediatric Dentists and Generated by ChatGPT: A Cross-Sectional Pilot Study.
兒童牙醫與 ChatGPT 所提供之診斷與治療計畫的比較評估:一項橫斷式前導研究
Cureus 2025-08-25
The paradox of creativity in generative AI: high performance, human-like bias, and limited differential evaluation.
生成式 AI 創造力的悖論:高效能、人類式偏誤與有限的差異化評估
Front Psychol 2025-08-25
2025 年有研究用「蛋任務」測試 ChatGPT-4o 的創造力,發現它雖然能產生比人類更多點子,但大多還是很普通,跟人類一樣有創意偏見,也不太會分辨哪些想法真的有創意。這代表生成式 AI 雖然能幫忙發想,但還是需要人類來挑選最有創意的點子。
相關文章PubMedDOI推理
A framework for robotic manipulation tasks based on multiple zero shot models.
基於多重 Zero Shot 模型的機器人操作任務框架
Sci Rep 2025-08-24
Performance of Advanced Artificial Intelligence Models in Pulp Therapy for Immature Permanent Teeth: A Comparison of ChatGPT-4 Omni, DeepSeek, and Gemini Advanced in Accuracy, Completeness, Response Time, and Readability.
先進人工智慧模型於未成熟恆牙牙髓治療之表現:ChatGPT-4 Omni、DeepSeek 與 Gemini Advanced 在準確性、完整性、回應時間及可讀性之比較
J Endod 2025-08-24
Development of a context-aware integrated training module based on large language models for continuous education in radiation protection.
基於大型語言模型的情境感知整合訓練模組於輻射防護持續教育之開發
Phys Med 2025-08-24
Evaluating large language models as graders of medical short answer questions: a comparative analysis with expert human graders.
將大型語言模型作為醫學簡答題評分者之評估:與專家人工評分者的比較分析
Med Educ Online 2025-08-24
Large language models underperform in European general surgery board examinations: a comparative study with experts and surgical residents.
大型語言模型在歐洲一般外科專科考試中的表現不佳:與專家及外科住院醫師的比較研究
BMC Med Educ 2025-08-24