Perceptions, acceptance and educational use of generative artificial intelligence among health professions students and educators: a rapid systematic review
This rapid systematic review of 20 studies involving nearly 9,553 health professions stakeholders reveals that while generative AI is generally accepted as a conditional learning aid for productivity and feedback, its integration requires robust AI literacy, verification skills, and ethical governance to address concerns regarding hallucinations, critical thinking erosion, and clinical accountability.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
Imagine the world of learning as a massive, bustling library where students and teachers are trying to figure out how to study for their future jobs as doctors, nurses, and healers. For a long time, the only tools in this library were textbooks, lectures, and human mentors. But recently, a new, incredibly fast, and chatty robot librarian has arrived. This robot, powered by something called "Generative Artificial Intelligence" (or GenAI for short), can write stories, summarize books, answer questions, and even explain complex medical cases in seconds. It's like having a super-smart tutor who never sleeps and knows almost everything. However, there's a catch: this robot sometimes makes things up, gets facts wrong, or gives advice that sounds perfect but is actually dangerous. Because the future of patient care depends on getting things right, the big question isn't just "Can this robot help us?" but "Can we trust it, and how do we use it without getting tricked?"
This paper is a "rapid systematic review," which is a fancy way of saying the authors acted like detectives who gathered every single clue (scientific studies) they could find between late 2022 and mid-2026 to answer that question. They looked at 20 different studies involving nearly 9,600 health students and teachers from around the world. Their mission was to see how these people felt about the new robot tutor: Did they love it? Did they fear it? And most importantly, how did they decide when to let the robot do the work and when to take over themselves?
The detectives found that the story isn't a simple "yes" or "no." Instead, it's more like a carefully calibrated dance. The students and teachers generally think the robot is a fantastic tool for brainstorming ideas, summarizing long texts, or getting quick explanations—like having a super-fast study buddy. They love how it can help them learn faster and be more productive. However, their trust is very conditional. It's as if they are willing to let the robot drive the car on a quiet, empty road (like studying for a quiz), but they absolutely refuse to let it drive on a busy highway with patients in the backseat (like making a real medical diagnosis or grading a final exam).
The main finding is that acceptance depends entirely on the "stakes." When the task is low-risk, like practicing a conversation or drafting an essay, people are happy to use the AI. But as soon as the task involves real-world consequences, like clinical reasoning or professional accountability, the trust drops. The paper suggests that while the robot is great at generating ideas, it is not yet ready to be the final authority. The biggest worries for everyone are that the robot might "hallucinate" (make up facts that sound real), that students might become too reliant on the tool to think for themselves, and that it's hard to tell if someone is submitting work without proper verification.
The authors point out that experience matters. People who have used the AI before or know how to ask it good questions tend to like it more, but even they remain cautious. They found that most people want clear rules: "You can use the robot for this, but not for that." They also want teachers to be the ones who double-check the robot's work. The paper explicitly argues against the idea that we should just ban the robot or let it run wild; instead, it suggests a "supervised augmentation" model. Think of it like a pilot and a co-pilot: the AI can help navigate and handle the controls, but a human must always be in the seat, ready to take over if the system glitches.
In short, the paper concludes that health professions are not rejecting this new technology, but they are learning to use it with a safety net. The robot is a powerful tool for learning, but it is not a replacement for human judgment. The future of medical education, according to these findings, will involve teaching students how to be "AI literate"—knowing how to use the tool, how to spot its mistakes, and when to trust their own brains over the machine's. The authors suggest that without these rules and without teaching students how to verify the robot's answers, we risk creating a generation of professionals who might rely too much on a tool that doesn't always tell the truth.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.