The impact of AI on medical students: How medical education is changing

AI in medical education is opening up new ways for students to research, learn and complete everyday academic tasks.

But as AI becomes increasingly embedded in academic and clinical life, questions about over-reliance, ‘cognitive offloading’ and the development of critical thinking skills are more relevant than ever.

Australian universities are now navigating where AI belongs in medical education — and what the impact of AI on medical students could mean for the next generation of doctors.

How are medical students using AI?

At the University of Sydney, Associate Professor Joanne Hart and colleagues examined how students in the Doctor of Medicine (MD) program were using generative AI tools in their research projects.

Since 2024, students have been permitted to use these tools, provided they comply with academic integrity policies and declare each use.

In 2024, 56% of students reported using AI in their research projects. By 2025, that figure had jumped to 81%.

“Our medical students report using AI thoughtfully, and a key finding was that they were careful to verify any AI outputs,” Associate Professor Hart told Insight+.

“They were using AI for many different research tasks, such as creating research timelines, transcribing interviews, translating documents, editing and proofreading. But they also used it to investigate — why use a particular statistical test, appraising ethical issues, brainstorming research gaps.”

AI can enhance learning – but it can’t replace it

For all its potential, experts caution that AI in medical education needs to be approached thoughtfully.   

As the Royal Australian College of General Practitioners (RACGP) notes, AI brings clinical risks — including ‘hallucinations’ and errors, embedded bias and data privacy concerns — that make human oversight essential.

“AI is excellent for things like reviewing and condensing large volumes of information, and that’s where it can offer enormous benefit in medical education and clinical practice,” says Dr Sean Stevens, a GP and Chair of the RACGP Digital Health and Innovation Special Interest Group.

“The trouble is, it’s so good that it can be easy to just fall back on it. Being able to critically appraise AI output is going to be an essential skill for future medical education.”

When AI starts to do the thinking

Dr Michael Tran, a GP and Senior Lecturer at the University of New South Wales, says there’s a deeper layer to the impact of AI on doctors in training to consider.

“AI works best in education when it acts as a scaffold that stretches thinking, rather than a substitute for it,” he says. “In practice, that means using it to interrogate your thinking, rather than simply retrieving an answer.”

Dr Tran notes that much of the concern about AI in medical education has focused on ‘metacognitive laziness’ — letting AI do the thinking students need to do themselves to develop clinical judgement and critical reasoning skills.

But the bigger concern, he says, is what he refers to as ‘epistemic automation’ — where students can start to conflate their own critical thinking capacity with the AI’s.

“It’s less about skipping the effort of reasoning, and more about gradually losing the internal benchmark for what good reasoning even feels like,” he says.

“This is because the fluent, confident output from AI tools can quietly substitute your own judgement, rather than sitting alongside it. Laziness is a choice you can reverse. Losing the capacity to tell your own reasoning from AI’s is much harder to walk back.”

AI in Australian universities: How are medical schools responding?

Medical school AI policies are increasingly shifting from restricting AI towards defining what responsible use looks like.

“The pace of technology is changing so rapidly that it is a real challenge for universities,” says Dr Stevens, who has been involved in helping medical schools develop AI education policies.

“The first thing is you need to have a commitment to AI education, and you need to make a policy. Student use of AI is always going to be moving faster than regulation can keep up with, so it’s a matter of trying to pre-empt things.”

However, he cautions against overregulating at the expense of preparing future doctors for AI.

“We do need regulation, and we do need to make sure people aren’t losing those essential skills. But every bit of regulation comes with issues, so it needs to be a balanced approach.”

Building foundational AI skills

Dr Stevens, Dr Tran and Associate Professor Hart all emphasise the need to equip medical students with foundational AI skills — from understanding how AI works to critically appraising outputs and protecting patient data privacy.

“Many of our students are very adept AI users, but there are also many who avoid it,” Associate Professor Hart says. “We need some base-level capabilities in prompting and data entry into AI, and they need to be able to critically evaluate AI outputs.

“It is imperative that they understand where the data is going, protecting privacy and instilling ethical, responsible use of the technology.”

Rethinking assessments and academic integrity

AI is also prompting some medical schools to rethink how they assess students and protect academic integrity, with some placing greater emphasis on in-person assessments and oral exams, where students can’t rely on AI.

“Trying to police or detect AI use item-by-item is a losing game,” Dr Tran says. “The more durable fix is redesigning what we actually reward, so that demonstrating your own reasoning is unavoidable rather than optional.

“This has to happen at every level of education, with educators changing how they teach, institutions changing how they assess, and colleges and regulators setting the standards that make those changes worth making.”

Preparing future doctors for AI

Preparing future doctors for AI means equipping them to use the technology responsibly while preserving the essential clinical skills needed for good medical practice.

“AI and technology-enhanced care aren’t optional extras doctors can opt out of — they’re an inevitability in how medicine will be practised,” Dr Tran says. “The question isn’t whether to engage with them but how to do it responsibly.”

While the tools doctors use may change, Dr Tran says the fundamentals of good clinical care remain.

“What develops good doctors remains the same as ever: deliberate, effortful engagement inside a trusted supervisory relationship,” he says.

“AI should extend that educator-learner relationship, but not replace it. In much the same way as we would not want a doctor’s only source of feedback to be a textbook, so too should we be careful about AI supplanting other valuable resources in learning.”

Key takeaways

• Medical students are already using AI for study and research, often with limited formal AI education.
• AI in medical education has clear potential benefits, including personalised learning and assisting with research tasks, but it must be used carefully. 
• Experts warn that over-reliance on AI could impact students’ foundational knowledge, critical thinking and clinical reasoning.
• Medical schools are adapting their policies, assessment and curricula to respond to AI — shifting from restricting AI to defining responsible use.
• Preparing future doctors for AI means developing AI literacy alongside the clinical and professional skills AI cannot replace.

FAQs

How are students using AI in medical education?

Research suggests many medical students are using AI tools for tasks like literature searches, writing and editing, transcription, translation and data visualisation. They also use it to support their learning, such as understanding why a particular statistical test is appropriate.

How are medical schools preparing future doctors for AI?

Some medical schools are now building dedicated AI education into their programs, but it varies. The Australian Medical Council’s national accreditation standards cover digital health but not AI specifically, so the approach to AI in Australian universities is up to each medical school.

How are medical schools handling AI and academic integrity?

Medical school AI policies are still developing. Some universities allow students to use AI in assignments as long as they declare each use and follow academic integrity policies. Other universities are shifting to more in-person, time-based assessments where AI can’t be used.

Can relying on AI affect clinical reasoning?

Yes, one concern about the impact of AI on medical students and doctors in training is that over-reliance on these tools could hinder the development of critical thinking and clinical reasoning skills. AI errors can also be harder for less experienced learners to recognise, making critical appraisal an essential skill.