Lecturers think students use AI far more than students report, study finds

University lecturers believe students rely on generative AI for their studies far more than students themselves say they do, according to a study by The University of Manchester and Universidad del Pais Vasco/Euskal Herriko Unibertsitatea in Spain.

On a seven-point scale, academic staff put students’ use of tools such as ChatGPT for academic work at an average of 5.23. Students rated their own use at 3.11.

The researchers say this points to a gap in understanding between the two groups just as universities are working out how to respond to the technology.

Students were more upbeat than staff about what generative AI could do for them, including making them more productive, helping them learn and cutting their workloads. Their main reported uses were writing text, getting answers for assignments, and translation or grammar correction.

Staff experience and knowledge of AI varied widely, leaving some poorly placed to assess how students used it.

How the study worked

The research looked at how academic staff and students viewed their own use of generative AI tools and how they believed it was being used across higher education.

Researchers questioned 86 students and 45 academic staff from a social sciences school at a research-intensive UK university. The questionnaires were carried out between April and June 2024.

The researchers accept that students may have under-reported their AI use. Even so, they say the gap suggests staff may have only limited knowledge of students’ use of these tools.

Lecturers can find AI use hard to spot, especially when they see only a finished assignment rather than how it was put together. Students may also use AI on their own, away from contact with teaching staff, so academics get only part of the picture.

Students more positive about the benefits

Students scored the technology higher than staff did on how far it was likely to help their academic performance, for example by deepening their subject knowledge, boosting productivity and lightening workloads. They were also more confident than staff that the right conditions and support were in place for them to use AI tools.

The researchers suggest that much of the immediate draw of AI tools may be their ability to generate and polish text. They argue, though, that universities should look at the many different ways students use it instead of assuming its part in academic work is simple.

Staff familiarity with AI differed considerably. Some 63 per cent of staff who responded had tried at least one generative AI tool during the six months prior to the survey, but some had little practical experience. Only 14 per cent paid for a subscription to an AI tool.

Call to move beyond suspicion

The authors argue that universities should stop treating AI mainly as a technical problem or a threat to academic integrity. Technical training or better AI literacy among students will not on its own solve the problems AI poses in higher education, they say. Universities instead need to look at how staff and students make sense of AI, what they expect of it and how those expectations influence their use of it.

Assuming students mainly use AI to get around assessment rules risks building a relationship founded on distrust, according to the researchers. While accepting that AI can be misused to the detriment of academic integrity, they say universities need a more nuanced view of what students actually do.

They suggest teaching and assessment should change in response to what generative AI can do, with staff and students given more support to explore appropriate educational uses. They also stress that the study is a snapshot of one institution at one point in time, that perceptions and patterns of use may shift as AI tools develop quickly, and that further research with larger samples is needed.

Dr Mark Carrigan, a senior lecturer in education at Manchester who co-wrote the study and has recently written a book on the subject, said: “There is a tendency for the debate about AI in universities to focus on the risks of cheating or the promise of transforming education, but the reality is more complex.

“Our findings suggest that staff and students have quite different perceptions of how AI is being used, making it difficult for universities to develop effective responses based on assumptions rather than a clear understanding of students’ experiences.”

He added: “We need to move beyond a climate of suspicion and focus on how these tools can be used appropriately to support learning, while rethinking teaching and assessment for an AI-enabled world.”