TL;DR
Turn your wind-down time into listening time
- Thousands of audiobooks, podcasts and originals
- Listen on your phone, tablet or Echo — also offline
- Cancel anytime
Experts speaking at the Association of Health Care Journalists Fall Summit said medical students’ use of AI is not automatically cheating and can be taught transparently. They also warned that reliance on AI could weaken clinical reasoning, and argued trainees need practice making decisions without digital tools.
Medical students who use artificial intelligence for assignments are not automatically cheating, speakers said at the Association of Health Care Journalists Fall Summit on October 15, while cautioning that AI must not replace independent clinical reasoning. The discussion focused on how schools can teach students to use the tools openly and responsibly while preserving skills needed when technology is unavailable or unreliable.
Graciela Gonzalez-Hernandez, a computational biomedicine researcher and education leader at Cedars-Sinai Medical Center, said AI use is not the same as plagiarism. She described an approach in which students show how they construct prompts, review the resulting text and work through it with an instructor. Her point was that making AI use part of classroom instruction is more productive than asking students to conceal it.
Speakers also raised the risk of de-skilling. Tina Hernandez-Boussard, a Stanford biomedical informatics professor, recounted a case in which medical residents helping at a cardiac ward in Uganda lost electricity and had no AI available to interpret electrocardiograms. She said they were not confident synthesizing the information to make patient-care decisions. The account was presented as a warning about training, not as a published study establishing how common the problem is.
Gonzalez-Hernandez agreed that schools should provide practice without technology, including paper-and-pencil scenarios. She also described a possible benefit: AI could prompt clinicians to ask questions they missed and help identify language in medical notes that may reflect uncertainty or bias. The speakers did not argue that AI should make clinical decisions independently; they discussed how it might support, and potentially complicate, physicians’ work.
Training Doctors Beyond the Screen
The debate matters because medical students are learning in a setting where AI tools can assist with writing, information review and clinical tasks, but physicians remain responsible for decisions affecting patients. If students depend on tools without learning how to evaluate their output or reason independently, they may be less prepared when systems fail, are unavailable or provide misleading results.
At the same time, banning or driving AI use underground could make it harder for educators to teach students how to disclose and critique their use of the technology. The speakers’ approach points to a practical tension for medical schools: teach AI skills directly, while testing whether students can still synthesize information and explain their reasoning without assistance.
The discussion also touched on fairness in care. Gonzalez-Hernandez said clinical notes can convey doubt in ways that may vary across patient groups, and suggested AI could flag questions a clinician has not asked. That is a proposed use, not evidence that AI reliably corrects bias. Any such benefit would depend on how systems are designed and evaluated, and on clinicians’ ability to challenge their suggestions.
As an affiliate, we earn on qualifying purchases.
A Classroom Debate About AI
The comments were made during a session at the Association of Health Care Journalists’ Fall Summit in Los Angeles. Gonzalez-Hernandez spoke from her work in computational biomedicine and education at Cedars-Sinai. The session was moderated by Eli Cahan, a physician and investigative journalist, and also included Hernandez-Boussard, a Stanford professor of biomedical informatics.
The discussion did not announce a new school policy, formal guidance or research findings. It addressed how educators might respond to student use that is already part of academic work. Gonzalez-Hernandez said she makes her program AI-friendly by setting expectations for prompt construction and discussing generated text with students, rather than treating AI use itself as misconduct.
“I’m withholding judgment until we see whether and what kind of clinical reasoning actually is relevant for doctors to do.”
— Eli Cahan, session moderator and physician
How Schools Will Set the Rules
The summit discussion did not establish a shared definition of permitted AI use across medical schools, nor did it set rules for disclosure, grading or assessment. It is also unclear which clinical reasoning tasks AI may perform accurately enough to change what students need to learn. The speakers offered examples and concerns, but the source report did not provide comparative research measuring whether AI use improves student performance or causes lasting loss of skill.
The Uganda account was described by Hernandez-Boussard as an example of the risks of dependence; details about the residents, the clinical circumstances and the outcome were not provided in the report. Similarly, the idea that AI might counter bias in notes was presented as a possibility, not a demonstrated effect. Whether such systems can do so safely and consistently remains unsettled.
Testing Skills With and Without AI
The speakers pointed toward training that makes AI use visible and includes exercises without technology. Gonzalez-Hernandez said schools should create scenarios in which trainees work with paper and pencil, while also teaching them to examine AI-generated material rather than accept it uncritically.
No specific follow-up policy or research milestone was announced at the session. The next practical steps for schools will depend on how they define acceptable assistance, assess students’ independent reasoning and evaluate whether AI suggestions improve or distort clinical judgment. The speakers’ central message was that education must address both sides of the issue: using AI competently and retaining the ability to practice without it.
Key Questions
Did the speakers say medical students can use AI without restrictions?
No. Gonzalez-Hernandez said AI use is not automatically plagiarism and described teaching students to use it openly. The session did not set a universal policy or say every use is acceptable; students still need to follow their school’s rules.
Why are educators concerned about AI use in medical training?
They are concerned that relying on AI could weaken students’ ability to interpret information and make decisions independently, particularly if tools or electricity are unavailable. The speakers called for practice without technology as well as instruction in AI use.
Can AI help reduce bias in medical care?
Gonzalez-Hernandez suggested AI could flag missed questions or patterns in clinical notes that may reflect bias. The report presents this as a potential benefit, not a proven effect, and does not establish that AI systems reliably reduce bias.
Did the summit announce new medical-school rules?
No. The session discussed possible teaching approaches, including reviewing students’ prompts and AI-generated text and practicing without digital tools. It did not announce formal rules or a common policy for medical schools.
Source: rss
Halloween Picks
halloween
As an affiliate, we earn on qualifying purchases.
