A global survey of 23,218 university students across 109 countries asked one simple question: what is it actually like to study in the age of AI? The answers are more complicated – and more human – than you might expect.
Seven in ten university students have tried ChatGPT – but where you study shapes whether you are likely to be one of them. Students in Latin America are the most likely to have used it (78%), while those in the Middle East and Africa are the least (66%). The gap is not huge, but it is real. And it tracks closely with something else: internet access. Regions where fewer people are online also show lower ChatGPT adoption among students. The diamond markers show internet penetration rates from the World Bank (2023) – hover over them to compare. North America appears low here only because very few North American students were captured in this survey.
Ask any student what they use ChatGPT for and the answer is almost always brainstorming – but that is where the similarity ends. Engineering and computing students reach for it when they are stuck on code or cramming for exams. Arts students use it to polish their writing or push through a creative block. The heatmap below shows just how differently the same tool gets used depending on what you study. Hover over any cell to see the exact score.
Using AI to study does not feel the way the headlines suggest. Students are not anxious or ashamed – those emotions barely show up. What dominates is curiosity, followed closely by feeling calm and happy. The bars to the right of centre are longer, heavier, and more numerous than those on the left. That is the story: this generation has largely made peace with AI as a study tool – even if a quiet unease still lingers in the background. Hover over any bar to see the exact score.
Here is the tension at the heart of AI in education: students genuinely believe ChatGPT makes them better learners – and they also worry it makes cheating too easy. These are not two different groups of students holding opposite views. It is largely the same students, holding both at once. The red dots show how strongly students agree ChatGPT boosts their knowledge and study performance. The blue diamonds show how strongly they fear it enables plagiarism, spreads false information, and does the thinking for them. Both clusters sit well above neutral. Use the dropdown to see whether your field of study changes the picture.
The robots-stealing-jobs narrative does not quite hold up here. When asked about AI’s impact on the labour market, students are far more likely to see opportunity than risk. They expect AI to create new jobs, lift productivity, and make remote work easier. Yes, they acknowledge it will demand new skills and fundamentally change the nature of work – but that reads less like fear and more like pragmatism. The one belief that barely clears neutral? That AI will simply reduce the number of jobs. Students, it seems, are already planning for a future with AI – not against it.
Data sources:
Ravšelj, D., Aristovnik, A., Kerzic, D., Tomaževic, N., & Umek, L. (2024). Higher education students’ early perceptions of ChatGPT – Global survey data [Data set]. Mendeley Data. https://doi.org/10.17632/gwhdkj5y5k.2
World Bank. (2024). Individuals using the Internet (% of population) [Data set]. World Bank Open Data. https://data.worldbank.org/indicator/IT.NET.USER.ZS
Survey conducted: October 2023 to February 2024 | n = 23,218 students | 109 countries and territories
Visualisations built with R using plotly, dplyr, tidyr, readxl, readr, forcats, scales.