Why are London-based students using AI more than the rest of the UK?

Now, I’d hedge a bet that there is nothing mysteriously intrinsic about being inside the M25 that spurs students to use AI.

London Higher has found that there is a 12 per cent point difference between London students’ use of AI and the rest of the UK based on this year’s Student Academic Experience Survey of around 10,000 full-time undergraduates.

But, pinpointing precisely why students in London are using AI at a higher rate than the rest of the country could help us understand AI-use across the sector.

(I should say here that it is important to note that when the research refers to the use of AI in the data, it refers to the university-sanctioned use of AI tools only. So, in short – nothing that would constitute cheating).

While there is research indicating marginal differences in the frequency of use of AI by provider type, we can’t really draw conclusions from this because London has a pretty diverse provider base, and it is unlikely to be provider type driving the 12 percentage point difference.

Contributing factors highlighted in the briefing include a high proportion of international students (London has the highest proportion of international students compared to any other area of the UK). What is interesting is that this data can be broken down further. The research showed that students from outside the EU use AI at more than double the rate of UK students, and students from the EU use AI at nearly double the rate of UK students.

So, it’s not necessarily a case of being inside that place near Slough that’s driving AI, but that that place near Slough has a higher concentration of international students. Of course, it could also be that students in the courses most frequently chosen by international applicants are likelier to include AI (there was 20 per cent gap between students in STEM subjects and students in the arts and humanities in AI-use).

There are also other student groups we need to consider. Students with caring responsibilities use AI more than students with no caring responsibilities (London has a higher percentage of student parents and students with caring responsibilities). Students who travel more use AI than students who travel less (in the absence of campus universities, sky-high rents, and a reliance on Transport for London, London students tend to travel more than in other cities). Students who do not undertake paid employment are less likely to use AI in their studies than all students who have jobs (students in London are more likely to have jobs).

When AI-use is considered by student groups which happen to be concentrated in London, rather than by geographical location alone, these insights do tell us what we have known for a while: students with time restraints and pressures are more likely to turn to AI tools to help them balance their studies.

The problem is that students do not always have equal access to AI tools, for instance, not all students can afford to pay for access to higher-spec tools that require subscriptions to use them. Additionally, not all students have the same level of skills and competence to know how to use AI in the permitted way. In Jisc’s digital skills survey released last week, sixteen per cent said they had received specific training on utilising AI tools.

It is not lost on me that the student groups concentrated in London – which has a higher frequency of student AI use – also overlap with widening participation groups. Digital poverty and differential access are still prevalent, and our danger is that some students manage their time pressures competently with AI while others struggle through.