Ten recommendations for the use of generative artificial intelligence into epidemiology teaching: Results from an expert workshop
Following a 2026 expert workshop, this paper presents ten evidence-informed recommendations organized into governance, competencies, and feedback domains to guide epidemiology instructors in responsibly integrating generative AI into teaching while preserving core scientific skills and addressing ethical concerns.
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
In the study of how diseases spread and how populations stay healthy, a field known as epidemiology, the core work has always relied on human judgment. Experts must look at data, spot patterns, and decide what those patterns mean for public health. This requires a deep understanding of cause and effect, the ability to spot errors in how information is collected, and the skill to explain complex findings clearly. For decades, teaching these skills meant guiding students through the hard work of thinking for themselves, ensuring they could build a solid foundation of knowledge before they ever tried to solve a problem. Now, a new kind of technology has arrived that can write text, analyze numbers, and generate answers almost instantly. This technology, called generative artificial intelligence, offers powerful tools that can help with learning, but it also carries a hidden danger. If students use these tools too early or too easily, they might skip the hard work of thinking, leaving them unable to solve problems on their own when the technology is not there. The question facing teachers today is not whether to use these tools, but how to use them without losing the very skills that make an expert an expert.
A group of experts from universities and research institutes across Germany, Norway, and Austria gathered to answer this question. They met for three days in a workshop to create a set of rules for teaching epidemiology in the age of artificial intelligence. These ten recommendations are designed to help instructors balance the benefits of new technology with the need to keep students sharp and capable. The group, which included experienced teachers, students, and specialists in artificial intelligence, decided that the goal is not to ban these tools, but to integrate them carefully so that students learn to use them responsibly while still mastering the basics of their field.
The first and most important step the experts recommend is to set clear rules. Every class needs a written policy that tells students exactly what they can and cannot do with artificial intelligence. This includes rules about what kind of data can be shared with these tools, since some information is private and must be protected. The rules must also explain how students should admit when they have used the technology, just as a scientist must admit when they used a specific method to get a result. If a student uses the tool to check their grammar or to get help understanding a difficult concept, they should say so. If they use it to write an entire assignment, that is usually forbidden. The teachers themselves must follow these same rules, showing students how to use the technology honestly and transparently.
Beyond setting rules, the experts emphasize that teachers must stay up to date. The technology changes so fast that a teacher who learned about it last year might already be behind. Instructors need regular training to understand how these tools work, what their limits are, and how to use them in a way that helps students learn rather than replacing the learning process. This training should cover not just how to type a question into the computer, but how to judge whether the answer the computer gives is correct. Teachers must also learn how to spot when a student is relying too much on the machine, a problem the researchers call "never-skilling," where a student never actually learns the skill because the machine did the work for them.
To prevent this, the experts suggest that students must first learn the basics without any help from artificial intelligence. Before they are allowed to use the tools to write code or analyze data, they must be able to do these things by hand or with standard software. They need to understand how to calculate a risk, how to spot a flaw in a study, and how to interpret a result on their own. Only after they have built this foundation should they be allowed to use the tools to check their work, refine their writing, or explore new ideas. This ensures that the technology acts as a helper, not a crutch. If a student cannot explain why they made a certain decision in their analysis, they have not truly learned the material, even if the final answer looks correct.
When it comes to testing what students have learned, the experts warn that old methods like take-home essays are no longer enough, because a student could simply ask the computer to write the essay. Instead, teachers should use different ways to test students, such as oral exams, in-person presentations, or supervised writing sessions where the use of artificial intelligence is not allowed. These methods force students to show their thinking and explain their reasoning in real time. The experts also advise against using software that claims to detect whether a student used artificial intelligence. These detection tools are often unreliable and can make mistakes, unfairly punishing students who did not violate academic integrity. It is better to rely on clear rules, honest disclosure, and human judgment to ensure fairness.
The recommendations also highlight the importance of teaching students how to talk to the computer effectively. This is known as prompt engineering, which is simply the skill of asking the right questions to get the best answer. A student who asks a vague question will get a vague answer, while a student who asks a specific, well-structured question can get a much better result. Teachers should guide students in this skill, showing them how to break down complex problems into smaller steps and how to verify the information the computer gives them. This includes checking facts, looking for errors, and making sure the computer has not made up sources or data that do not exist.
Finally, the experts urge teachers to share what they learn with each other. Because this is a new field, no one has all the answers yet. Teachers should try different approaches, see what works and what does not, and then tell their colleagues about it. This could mean sharing a lesson plan that worked well, or admitting that a certain tool caused more problems than it solved. By working together and learning from each other's experiences, the entire field of epidemiology can adapt to this new technology in a way that keeps students safe, skilled, and ready for the future. The ultimate goal is to produce epidemiologists who can use these powerful tools to do better work, but who also possess the critical thinking skills to know when to trust the machine and when to trust their own judgment.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.