AI in ophthalmology research: Ethical implications for medical students

Over the past decade, earning an ophthalmology residency spot has become more difficult and competitive.¹ Between 2021 and 2024, the overall match rate declined from 74% to 66% and the number of residency applicants grew by 15%, despite an increase in available residency spots.^2,3 This caused a substantial increase in unmatched applicants and heightened competition for each position in each upcoming year. In 2022, another major change occurred that impacted the residency application process. USMLE Step 1 exams converted to pass/fail, as opposed to a scoring system.^4,5 This led to a greater focus on USMLE Step 2 scores, letters of recommendation, and research productivity as differentiating objective measures.

Medical students face several challenges when they are trying to obtain ophthalmology research opportunities, namely a limited exposure to ophthalmology curricula, lack of structured mentorship, and unequal access to resources.⁶ The emergence of artificial intelligence (AI) tools such as large language models (LLMs) and imaging algorithms have helped to bridge these challenges for medical students. An increasingly popular debate is the balance between these benefits with the limitations and the ethical considerations of AI.

Applications of AI in research

AI tools are widely used across the spectrum of research, including within the ophthalmology realm. Medical students may be utilizing them to compensate for gaps in experience, mentorship, and time constraints.^7,8 Khalifa et al. outlines that AI supports academic writing and research through idea generation and research design, improving content and structuring, supporting literature review and synthesis, enhancing data management and analysis, supporting editing and publishing, and assisting in communication and ethical compliance.⁹

Additionally, AI can adopt a mentorship role, successfully guiding students across various aspects of medical training, including career planning, through a personalized mentorship model.¹⁰ Generative LLMs, including ChatGPT and OpenEvidence, process large volumes of medical literature rapidly, making it easier to identify relevant studies and synthesize evidence.^11,12 Human processing can be slower, resulting in research that soon becomes outdated. In clinical research, the use of AI in ophthalmology is both helped and harmed by our reliance on imaging modalities (eg, optical coherence tomography (OCT) and fundus photography).¹³ Analysis of these technologies has been successfully completed by AI, with performance approaching that of humans.¹⁴

Aside from analysis of these specific modalities, AI can be utilized generally to analyze datasets or to provide recommendations for meaningful analysis.¹⁵ AI technologies are gifted in recognizing patterns that may be overlooked and are able to assist in most parts of the research process. However, its usage still requires human oversight and contextual interpretation.¹⁶

Ethical considerations

The benefits of utilizing AI in conducting scholarly work are not without significant ethical considerations that must be addressed.¹⁷ The use of generative models poses key ethical challenges that include academic integrity and plagiarism, data privacy and consent, bias and equity, transparency and authorship, as well as educational impact.¹⁸

Overreliance on AI-generated content risks compromising academic integrity. This leads to issues of plagiarism or the inclusion of fabricated or “hallucinated” references and content.¹⁹ Several studies over the past few years have found that ChatGPT, among other LLMs, appeared to utilize a predictive process rather than an accurate one, resulting in the generation of false citations and references.²⁰

Similarly, the use of patients’ sensitive imaging data and health information by generative AI brings about the risk of misuse in addition to issues of privacy and consent. The potential for bias in AI algorithms is a major concern, which can inadvertently result in disparities in care recommendations and worsen existing healthcare inequities.²¹

Effects on learning and critical skills

In addition, questions have been raised about whether the use of AI tools to generate medical literature is diminishing the human element of research.The process of conducting research traditionally involves a certain level of critical thinking and analytical skills, which are developed and refined over continued use. These are key competencies that medical students are expected to cultivate, and they often seek research opportunities outside the formal medical school curriculum to further their research education.

However, the overuse of AI tools risks turning research into a mechanical process that focuses on productivity and output generation. In addition to limiting student learning opportunities, this trend harms the quality of scientific work. In addition, allegations of the use of AI can be difficult to confirm and open the door to false allegations especially when utilizing so called AI detectors which have notoriously poor sensitivity and specificity.

AI literacy in medical education

Nevertheless, as AI becomes increasingly ubiquitous in everyday life and more heavily used in healthcare and research, the need for AI literacy is paramount amongst medical students and medical education. While studies have found that medical students feel comfortable practically using AI, many report a lack of technical understanding of AI and a desire for AI literacy training in medical education.^22,23 Many medical schools across the world have begun to address this gap by integrating AI literacy training into medical school education and through development of meaningful curriculum to enhance student understanding of AI, how to utilize this tool, and its clinical and ethical implications.^24,25

The concern still remains that medical students may become too reliant on these emerging technologies as a crutch for attaining medical knowledge, ophthalmology research pursuits to hit research “quotas”, and clinical decision making when training in the ophthalmology setting. It should be emphasized that artificial intelligence must be used as a complement to medical education and clinical or research practice as opposed to a replacement. AI literacy training should not only educate students on how to utilize and critically appraise AI responses, but also imbue students with an ethical awareness of the future implications of AI and how to prevent its misuse or over-reliance.

In conclusion, AI holds much promise in medicine and the field of ophthalmology. With the rapidly changing landscape of AI, its integration in ophthalmology and research has implications for the clinical practice of current and future physicians. Especially for medical students who are inundated with the use and knowledge of AI on a regular basis, the importance of balancing practices of responsibility and transparency when utilizing AI are more important than ever.

As it pertains to research and the growing pressure for medical students to attain high quantities of publications especially for competitive specialties, such as ophthalmology, and limited residency spots, there is a need to establish strict guidelines to safeguard the integrity of research.

Rohini Chahal, BSA; Nicole Koulov, BS; Sophie Saland, BA; and Amna Ali, BA, are students at the John P. and Kathrine G. McGovern Medical School at The University of Texas Health Science Center at Houston (UTHealth), Houston, Texas.
Correspondence: [email protected]

Andrew G. Lee, MD, is affiliated with Baylor College of Medicine and Blanton Eye Institute, Houston Methodist Hospital, University of Texas MD Anderson Cancer Center, in Houston, Texas; Departments of Ophthalmology, Neurology, and Neurosurgery, Weill Cornell Medicine, New York, New York; Department of Ophthalmology, University of Texas Medical Branch, Galveston, Texas; and the Department of Ophthalmology, The University of Iowa Hospitals and Clinics, Iowa City, Iowa.

The authors have no relevant disclosures to declare.

References

1. Rasendran C, Murali S, Kanagasegar N, Kapadia M, Lass J, Ohsie-Bajor L. An analysis of medical student PubMed-indexed research productivity and factors associated with matching at top-ranked ophthalmology residency programs. Cureus. 2024;16(1):e52824. doi:10.7759/cureus.52824

2. Ophthalmology residency match summary report 2021. April 12, 2021. Accessed October 29, 2025. https://aupo.org/sites/default/files/2021-04/Ophthalmology Residency Match Summary Report 2021.pdf

3. Ophthalmology residency match summary report 2024. April 26, 2024. Accessed October 29, 2025.https://aupo.org/sites/default/files/2024-04/2024Ophthalmology-Residency-Match-Report.pdf

4. Cohen SA, Pershing S. Relative importance of applicant characteristics in ophthalmology residency interview selection: a survey of program directors. J Acad Ophthalmol (2017). 2022;14(2):e246-e256. doi:10.1055/s-0042-1756122

5. Boyd LR, Lin LO, Janis JE. Matching into integrated plastic surgery: the impact of USMLE Step 1 transition to pass/fail: did the 2019 predictions come true in 2024? Plast Reconstr Surg Glob Open. 2025;13(1):e6417. doi:10.1097/GOX.0000000000006417

6. Liao J, Wright RR, Vora GK. The decline of basic ophthalmology in general medical education: a scoping review and recommended potential solutions. J Med Educ Curric Dev. 2024;11:23821205241245635. doi:10.1177/23821205241245635

7. Liu DS, Sawyer J, Luna A, et al. Perceptions of US medical students on artificial intelligence in medicine: mixed methods survey study. JMIR Med Educ. 2022;8(4):e38325. doi:10.2196/38325

8. Sami A, Tanveer F, Sajwani K, et al. Medical students' attitudes toward AI in education: perception, effectiveness, and its credibility. BMC Med Educ. 2025;25(1):82. doi:10.1186/s12909-025-06704-y

9. Khalifa M, Albadawy M. Using artificial intelligence in academic writing and research: an essential productivity tool. Comput Methods Programs Biomed Update. 2024;5:100145. doi:10.1016/j.cmpbup.2024.100145

10. Çetinkaya L. Redefining mentorship in medical education with artificial intelligence: a Delphi study on the feasibility and implications. Teach Learn Med. doi:10.1080/10401334.2025.2521001

11. Ruan M, Fan J, Liu M, Meng Z, Zhang X, Zhang C. Artificial intelligence for the science of evidence synthesis: how good are AI-powered tools for automatic literature screening? BMC Med Res Methodol. 2025;25(1):199. doi:10.1186/s12874-025-02644-9

12. Thirunavukarasu AJ, Ting DSJ, Elangovan K, Gutierrez L, Tan TF, Ting DSW. Large language models in medicine. Nat Med. 2023;29(8):1930-1940. doi:10.1038/s41591-023-02448-8

13. Labib KM, Ghumman H, Jain S, Jarstad JS. A review of the utility and limitations of artificial intelligence in retinal disorders and pediatric ophthalmology. Cureus. 2024;16(10):e71063. doi:10.7759/cureus.71063

14. Grzybowski A, Jin K, Zhou J, et al. Retina fundus photograph-based artificial intelligence algorithms in medicine: a systematic review. Ophthalmol Ther. 2024;13(8):2125-2149. doi:10.1007/s40123-024-00981-4

15. Keykha A, Behravesh S, Ghaemi F. ChatGPT and medical research: a meta-synthesis of opportunities and challenges. J Adv Med Educ Prof. 2024;12(3):135-147. doi:10.30476/JAMP.2024.101068.1910

16. Adel A, Alani N. Can generative AI reliably synthesise literature? Exploring hallucination issues in ChatGPT. AI Soc. 2025. doi:10.1007/s00146-025-02406-7

17. Yu S; Lee SS; Hwang H. The ethics of using artificial intelligence in medical research. Kosin Med J. 2024;39(4):229-237. doi:10.7180/kmj.24.140

18. Gracey LE, Nambudiri VE, Rotemberg V, Maderal AD. Challenges and opportunities for AI in dermatology residency. JAMA Dermatol. doi:10.1001/jamadermatol.2025.3366

19. Walters WH, Wilder EI. Fabrication and errors in the bibliographic citations generated by ChatGPT. Sci Rep. 2023;13(1):14045. doi:10.1038/s41598-023-41032-5

20. Day T. A preliminary investigation of fake peer-reviewed citations and references generated by ChatGPT. Prof Geogr. 2023;75(6):1024-1027. doi:10.1080/00330124.2023.2190373

21. Koirala P, Thongprayoon C, Miao J, et al. Evaluating AI performance in nephrology triage and subspecialty referrals. Sci Rep. 2025;15(1):3455. doi:10.1038/s41598-025-88074-5

22. Laupichler MC, Aster A, Meyerheim M, Raupach T, Mergen M. Medical students' AI literacy and attitudes towards AI: a cross-sectional two-center study using pre-validated assessment instruments. BMC Med Educ. 2024;24(1):401. doi:10.1186/s12909-024-05400-7

23. Wood EA, Ange BL, Miller DD. Are we ready to integrate artificial intelligence literacy into medical school curriculum: students and faculty survey. J Med Educ Curric Dev. 2021;8:23821205211024078. doi:10.1177/23821205211024078

24. Laupichler MC, Hadizadeh DR, Wintergerst MWM, et al. Effect of a flipped classroom course to foster medical students' AI literacy with a focus on medical imaging: a single group pre-and post-test study. BMC Med Educ. 2022;22(1):803. doi:10.1186/s12909-022-03866-x

25. Singla R, Pupic N, Ghaffarizadeh SA, et al. Developing a Canadian artificial intelligence medical curriculum using a Delphi study. NPJ Digit Med. 2024;7(1):323. doi:10.1038/s41746-024-01307-1