The use of large language models (LLMs) is soaring among students worldwide thanks to the tools’ growing ability to solve problems and generate text and code. In a 2026 survey of 1,054 UK undergraduate students by the Higher Education Policy Institute in Oxford, UK, roughly 94% said that they use generative artificial-intelligence tools to help them with assessed work, and 12% directly inserted AI-generated text into their coursework¹. In a study published in May, from survey data of more than 95,000 students at 20 US universities, the authors estimated that 9% of students used AI on their coursework, despite knowing that this broke the rules².

Educators worldwide are seeing suspicious signs of AI use — from hallucinated references to punctuation styles that are typical of AI tools — in take-home coursework and online examinations, fuelling worries that the technology is giving some students an unfair advantage over those who don’t use it. At one point, people were submitting essays in which “50% of the reference literature did not exist in real life”, says Nikita Bezrukov, who teaches linguistics and communication at the Massachusetts Institute of Technology in Cambridge. At the same time, many professors are increasingly recognizing the benefits and inevitability of AI use in students’ education and future careers, and see a need to teach young people to use the technology wisely.

These realizations are forcing many educators to rethink conventional assessment approaches, both to accommodate AI as a tool and detect its misuse. Nature’s careers team spoke to educators around the globe about some of the approaches they and their institutions are taking to revamp assessment.

That said, “I don’t think anybody has a ‘right’ answer right now”, says computer scientist Nicholas Mattei at Tulane University in New Orleans, Louisiana. “Things are changing pretty rapidly.”

AI-based assignments

“For an assignment in my history-of-technology class, students have to identify sources, use three LLMs to summarize them and critique what the models say. Then, they make infographics and write an essay, for which AI use is allowed. They get extra points if they notice when the infographics don’t accurately reflect the sources or if the LLMs hallucinate references. The goal is to prompt more critical reflection on what comes out of LLMs.” — Nicholas Mattei.

“Instead of writing an essay about Adam Smith’s 1776 economics book The Wealth of Nations, which is what I used to ask students to do, I got them to build a virtual representation of his theory using AI agents acting as traders in a marketplace. We then explored various potential situations — such as what happens if people become untrustworthy. Students analysed the transcripts produced by the agents and wrote about it. It was easy to spot whether students asked AI to do all these steps in one go because, for instance, it just made stuff up rather than quoting from the transcripts.” — Daniel Silver, sociologist at the University of Toronto Scarborough, Canada.

On allowing AI use

“The exam for my web-development course is open book and open internet, and AI is part of the internet. But I’m assessing a skill that AI doesn’t have — how to build good, sustainable web applications. I ask inherently open questions that have multiple answers, such as “This website doesn’t work, why?” and “Why is this website slow?” If you put this into AI chatbots, they just answer questions without sound methodology, and not even always accurately.” — Ruben Verborgh, computer scientist at Ghent University, Belgium.

“I cannot trust that students aren’t using AI, so by default, I’m grading the student–AI team. But the task that I assign isn’t something that can be done with only ChatGPT. I provide screenshots of a specific kind of analysis being performed on data of neurological activity, and ask students to write a narrative about it. If students just paste the assignment prompt into a chatbot and use the AI’s output verbatim, the answer often has nothing to do with what’s happening in the screenshot, and it shows.” — Etienne Roesch, statistician and cognitive scientist at the University of Reading, UK.

Show your work

“For students — especially PhD students — using AI in research and academic writing, I ask them to share their original interaction logs with AI tools. The purpose is not simply to monitor students, but to help me to evaluate whether they were deeply involved in the intellectual work, what kinds of knowledge they expanded through AI and whether the tool was used to support their thinking rather than to replace it.” — Yanjun Shen, researcher in ecological and engineering geology at Chang’an University in Xi’an, China.

“Instead of submitting coursework as PDFs, my students need to use software that has a revision history, such as Google Docs. This enables me to look at their earlier versions, so I can see how and whether they used AI in the process. I don’t care if, for instance, someone with English as a second language uses it to help with grammar or polishing of the text, but using AI to generate text wholesale is prohibited.” — Nikita Bezrukov.

Less take-home coursework

“For one of the courses I teach, I’ve changed the assignments so that fewer are written at home. I used to ask students to make two infographics, but after seeing a lot of AI-generated material, I replaced this with an in-class pitch about a higher-education policy or programme and later, a full presentation. We also do more impromptu speeches and in-class writing assignments than we used to.” — Kristina Ruiz-Mesa, communications instructor at California State University in Los Angeles.

“Although I had moved away from conventional examinations, this spring, I went back to written, in-class exams. I also started having students write essays in class three times in a semester. They can bring notes and books, but they have to write in an examination booklet, with no phones or laptops.” — Jennifer Sessions, historian at the University of Virginia in Charlottesville.

Ungraded assignments

“For take-home assignments in data analytics, I give only pass or fail grades. It’s not wrong if they use AI to generate code so that they can save time for more-interesting analyses. However, if they substitute their own thinking with AI, they might pass the coursework but will probably struggle in the final exam, in which only basic calculators are allowed. By removing the grading, we encourage true learning: we’ve never had as many students achieving full-score grades in the exams as we did this year.” — Natalia Sidarova, computer scientist at Eindhoven University of Technology, the Netherlands.

“In my department, we allow AI for take-home assignments, but some colleagues reduce the weightage of these and assess the learned skills in an offline quiz. Some of the quiz questions directly relate to the assignments, increasing the probability of students succeeding if they did the assignment themselves. We also gamify coursework with a leaderboard that compares students’ performance to motivate them to work harder.” — Mausam, computer scientist at the Indian Institute of Technology Delhi, India.

Shifting to oral examinations

“I get students to write a lot of Python code to interpret chemical data. Seventy per cent of the points awarded are for how well the code works; the remaining 30% depends on a brief one-to-one oral exam in which I ask them how specific lines of code work. I don’t care if ChatGPT told them a line; they need to understand and be responsible for every line of code.” — Gianmarc Grazioli, computational chemist at San José State University, California.

“The written portion of the first piece of coursework I assign to students now counts for a very small percentage of the final grade; what counts the most is a conversation with me in person afterwards to discuss and further develop the material. For later exams, if students can’t demonstrate authorship and mastery during the discussion with me, it caps their grade at a C.” — Thea Goldring, lecturer in Princeton University’s Princeton Writing Program, New Jersey.

AI-detection tools

“For the assignments that students submit for my online class, I use an AI-detection tool provided by the university, which can provide some level of corroboration to my gut instinct. Another clue is hallucinated references or when students’ first drafts and final essays are drastically different. I try to make a reasonable assessment from several pieces of information.” — Jacqueline Evans, psychology researcher at Florida International University in Miami.