Volume 16, Issue 6 And S7 (Artificial Intelligence- In Press 2026)
J Research Health 2026, 16(6 And S7): 3-3 |
Back to browse issues page
Ethics code: Review article
Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
Mirnezami S R, Keykha A. Unraveling the Agency Spectrum: A Mixed‑Methods Research Synthesis of Opportunities and Challenges of AI Agents in Medical Education. , J Research Health. 2026; 16 (6 and S7) :3-3
URL: http://jrh.gmu.ac.ir/article-1-3189-en.html
URL: http://jrh.gmu.ac.ir/article-1-3189-en.html
1- The Research Institute for Science, Technology, and Industry Policy (RISTIP) Sharif University of Technology, Tehran, Iran.
2- Department of Educational Administration, School of Psychology and Education, Kharazmi University, Karaj, Iran ,ahmadkeykha8984@gmail.com
2- Department of Educational Administration, School of Psychology and Education, Kharazmi University, Karaj, Iran ,
Abstract: (25 Views)
Background: It is a matter of great social and scientific interest to understand the involvement of medical educational research in this change. Yet in the literature, basic chatbots are often confused with autonomous systems, thus ignoring the opportunities and risks of greater agency.
Methods: The present study was a mix-method research synthesis of English-language peer-reviewed journals published between March 2015 and March 2026. Data sources searched were Science Direct, Springer, ERIC, Emerald, SAGE Journals, Wiley online library, Pub Med, google scholar, and ArXiv. The initial screening identified 180 records, but after removing 12 duplicate items 168 records were screened for inclusion in the review, ultimately 24 were included. The quality of the included studies was assessed using QuADS tool. The methodological approach consisted of two layers, a qualitative content analysis on the identified articles based on existing framework (Vaismoradi, 2016), to identify emerging themes of the opportunities and challenges of AI use, and the novel Agency Spectrum Framework to code each of the identified concepts into General/Basic AI, Agent-Enhanced AI, and Exclusive Agent AI, based on four dimensions of autonomy, memory, planning, and proactiveness.
Results: Eight areas of opportunities and challenges for AI are uncovered. Spectrum disaggregation reveals an important asymmetry of opportunities and challenges: the most pedagogically beneficial uses of AI- automation, errors, rudimentary personalization and simulation-can all be achieved with general AI, but the benefits are greatly enhanced by contextual memory and adaptability in the form of adaptive feedback and real-time adjustment. A smaller, distinct set of three-dynamic reasoning loops, multiagent diagnostic coordination, and autonomous self-correction-can only be realized with fully autonomous agents. While all challenges also occur on the general level, a further eight specific challenges-cascading misinformation, accountability gaps, and bias reinforcement-will intensify with greater agency and peak with multiagent autonomous systems.
Conclusion: AI agents are already powerful and potentially revolutionary technology. AI's distribution of risks and benefits will be uneven – the range of benefits will be accessible at relatively low levels, whereas the range of dangers will increase substantially as increased agent autonomy increases. Our proposal calls for proportionate governance – perhaps in the form of enhanced transparency and human oversight.
Methods: The present study was a mix-method research synthesis of English-language peer-reviewed journals published between March 2015 and March 2026. Data sources searched were Science Direct, Springer, ERIC, Emerald, SAGE Journals, Wiley online library, Pub Med, google scholar, and ArXiv. The initial screening identified 180 records, but after removing 12 duplicate items 168 records were screened for inclusion in the review, ultimately 24 were included. The quality of the included studies was assessed using QuADS tool. The methodological approach consisted of two layers, a qualitative content analysis on the identified articles based on existing framework (Vaismoradi, 2016), to identify emerging themes of the opportunities and challenges of AI use, and the novel Agency Spectrum Framework to code each of the identified concepts into General/Basic AI, Agent-Enhanced AI, and Exclusive Agent AI, based on four dimensions of autonomy, memory, planning, and proactiveness.
Results: Eight areas of opportunities and challenges for AI are uncovered. Spectrum disaggregation reveals an important asymmetry of opportunities and challenges: the most pedagogically beneficial uses of AI- automation, errors, rudimentary personalization and simulation-can all be achieved with general AI, but the benefits are greatly enhanced by contextual memory and adaptability in the form of adaptive feedback and real-time adjustment. A smaller, distinct set of three-dynamic reasoning loops, multiagent diagnostic coordination, and autonomous self-correction-can only be realized with fully autonomous agents. While all challenges also occur on the general level, a further eight specific challenges-cascading misinformation, accountability gaps, and bias reinforcement-will intensify with greater agency and peak with multiagent autonomous systems.
Conclusion: AI agents are already powerful and potentially revolutionary technology. AI's distribution of risks and benefits will be uneven – the range of benefits will be accessible at relatively low levels, whereas the range of dangers will increase substantially as increased agent autonomy increases. Our proposal calls for proportionate governance – perhaps in the form of enhanced transparency and human oversight.
Keywords: Artificial Intelligence, Clinical Decision-Making, Educational Technology, Medical Education
Type of Study: Review Article |
Subject:
● Artificial Intelligence
Received: 2026/08/11 | Accepted: 2026/10/13 | Published: 2026/11/11
Received: 2026/08/11 | Accepted: 2026/10/13 | Published: 2026/11/11
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |
This is an open access article distributed under the terms of the Creative Commons Attribution License (CC-By-NC), which permits use, distribution, and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes.
Contact Information
