Clinical Report: AI in Medicine: 'Not Everything is What it Appears to Be'

Background

The integration of AI into healthcare has the potential to enhance diagnostic accuracy and clinical decision-making. However, as AI systems become more capable, it is crucial for physicians to maintain their clinical judgment to determine the appropriate use of these technologies.

Data Highlights

No numerical data available in the source material.

Key Findings

• AI's performance on medical examination questions has improved significantly, with recent models scoring 90% or higher.
• A randomized trial showed no significant improvement in diagnostic reasoning when physicians used GPT-4 alongside conventional resources (76% vs 74%).
• GPT-4 used alone scored 16% higher than physicians using conventional resources.
• In a second trial, physicians using GPT-4 scored 6.5% higher on complex patient-management tasks compared to those using conventional resources alone.
• 20% of frontier AI models recognized clinical concerns that none of the general internal medicine physicians identified in a real-world consultation study.
• Successful AI implementation requires attention to change management, integration, and workflow.

Clinical Implications

Physicians must critically assess AI outputs and integrate them into their clinical reasoning processes.

Conclusion

The effective use of AI in medicine hinges on the balance between technological capabilities and the indispensable role of clinical judgment.

Related Resources & Content

1. WHO, WHO, 2024 -- Ethics and governance of artificial intelligence for health: Guidance on large multi-modal models
2. HHS/ONC, HHS, 2024 -- HTI-1 Final Rule - ONC - Office of the National Coordinator for Health Information Technology
3. FDA, FDA, 2026 -- Artificial Intelligence-Enabled Medical Devices | FDA
4. National Academies, National Academies Press, 2025 -- An Artificial Intelligence Code of Conduct for Health and Medicine: Essential Guidance for Aligned Action
5. Nature Medicine, Nature Medicine, 2026 -- AI-based triage and decision support in mammography and digital tomosynthesis for breast cancer screening: a paired, noninferiority trial
6. aace endocrine ai — AI in endocrinology: Promises, risks, and responsibilities
7. aace endocrine ai — From skeptic to user: How AI earned a place in my practice
8. Stat News — Using AI in addiction medicine could be particularly risky
9. asco ai in oncology — AI in Oncology: From Diagnosis to Human Connection
10. AI in endocrinology: Promises, risks, and responsibilities
11. From skeptic to user: How AI earned a place in my practice
12. Using AI in addiction medicine could be particularly risky
13. Ethics and governance of artificial intelligence for health: Guidance on large multi-modal models
14. HTI-1 Final Rule - ONC - Office of the National Coordinator for Health Information Technology
15. Artificial Intelligence-Enabled Medical Devices | FDA
16. An Artificial Intelligence Code of Conduct for Health and Medicine: Essential Guidance for Aligned Action | The National Academies Press
17. AI-based triage and decision support in mammography and digital tomosynthesis for breast cancer screening: a paired, noninferiority trial | Nature Medicine
18. Real-Time Artificial Intelligence Versus Standard Colonoscopy in the Early Detection of Colorectal Cancer: A Systematic Review and Meta-Analysis - PMC
19. AI-enabled electrocardiography alert intervention and all-cause mortality: a pragmatic randomized clinical trial | Nature Medicine
20. Large language model influence on diagnostic reasoning: a randomized clinical trial. | PSNet
21. Evaluation and mitigation of the limitations of large language models in clinical decision-making. | PSNet