AI-ncriminate Yourself: The Results of The Saint's Student AI Survey Are In

Percentages are rounded to the nearest whole number. A full breakdown of the anonymous survey results is available here.

In an anonymous survey shared by The Saint, 70 per cent of the students surveyed said that they use artificial intelligence (AI) for university work. Of that AI-philic 70 per cent, a third used AI at least weekly; just under a fifth use it ‘near-daily’. 

I’m guessing you aren’t too surprised. You could figure that out from a snoop around the Main Library basement. I wanted to know more with my survey: what, exactly, are people using AI for? What do they think it should be used for? Should you use it to message a flatmate? A family member? 

I started with the most judgy questions: best to get them over with. Did AI-philes feel ‘immoral’ about using AI for coursework? Out of 104 respondents, 53 per cent of them did, to some extent. What about AI-phobes, those who put ‘never’ when asked if they use AI? Over 70 per cent of them feel ‘morally superior’ for not using AI.

A quick disclaimer: this survey is not the last word on AI in St Andrews. We could only cajole around one per cent of the undergraduate body. Nearly all the respondents were undergraduates. (Why do postgrads spurn The Saint? Is it their crushing workloads? Or are they just unable to appreciate the delights of student journalism?). The nationalities, at least, were representative of St Andrews — around a third of respondents were Scottish, and just under half were international. That reflects the University's actual makeup.

Over two-thirds of AI-philes used it for planning coursework, around a third used it for ‘editing’ essays, and one in eighteen used it for ‘completely writing’ assignments. (Just to reiterate: the survey was anonymous. I don’t know who these people are, Dame Sally.) 

What about tutorial readings? How many sleep-deprived students heed the siren call of the JSTOR AI pop-up? 44 per cent of AI-philes used it for ‘helping understand’ readings they had done, while 51 per cent used it for ‘summarising’ assigned readings they had not done. One in five AI-philes use it when ‘revising’ for exams, and around one in fifteen admitted to using it for ‘cheating on exams’.

When I asked the AI-philes if they tried to justify using AI for uni work, there were some interesting write-in answers. In some answers, you could feel the guilt seeping from the text: “I promised myself not to do it again.” Others were less repentant: “I’m working smarter not harder.” One brave soul opted for, “My peers do it.” Mostly I got variations of ‘It’s OK because I only use AI for planning essays, not actually writing them.’ 

This was a crucial distinction for the AI-philes: when I asked them how people should use AI, 78 per cent opposed AI written essays (even if the AI use is cited) but 74 per cent were OK with using AI for ‘planning/researching’ essays. It’s like Google — right? Right? 

The AI-phobes, on their part, were strongly opposed to using AI in both planning and writing coursework. (74 per cent and 94 per cent respectively.) They didn’t present such a united front when I asked about tutorial readings, however. While 90 per cent were against using AI for summarising readings you haven’t bothered doing, only 45 per cent opposed using AI to help decipher readings you did do. AI-phobes were in agreement about one more thing: a good 55 per cent didn't think abstaining from AI would hurt their academic performance. 

One AI-phobe, Diane Buffet-Mogel, DMed me after taking the survey about exactly that. She wasn’t worried that avoiding AI would hurt her grades: "It’s hard work being smarter, cooler, and funnier than everyone else," she messaged.

So that’s the deal with using AI for uni work. But what do these brave respondents think about the other messier conflict: using AI in your personal life? 

What about, for instance, having an AI-generated Linkedin bio? 82 per cent of AI-philes were fine with that, compared to only 36 per cent of AI-phobes (another 32 per cent of AI-phobes weren’t sure.)

Now, what about AI-generated Tinder bios? If it’s aimed at attracting men, 51 per cent of AI-philes were okay with it; if it’s aimed at attracting women, that approval falls to 47 per cent. In contrast, only 19 per cent of AI-phobes supported robotically-written bios aimed at either men or women. 

Computer-crafting a DM to a flatmate? Thumbs up from two-thirds of the AI-philes and a wiggly sideways thumb from the AI-phobes (half said it was okay, a quarter weren’t sure, a quarter were against).

AI-drafted DMs to your mum? The wiggly thumb is given by the AI-philes this time (they’re pretty evenly split) while a stern thumbs down from the AI-phobes (only 29 per cent approved.)

Procedurally generated DMs to your romantic partner gets the rare double-thumbs down: a majority of both AI-philes and phobes were opposed. Love, excluding the maternal variety, isn’t dead yet.

What about when other people in your life use AI for their education? Say, your elected representative? Or your doctor? Or your pilot? The AI-philes didn’t feel too strongly. For each of the three professions, a majority responded it ‘depends how (much) they used it.’ 

The AI-phobes gave very different results. Over 70 per cent said they definitely wouldn’t want their doctor, pilot, or elected representative to have used AI during their education. Interestingly, out of the three, the profession that drew the strongest emotional reaction (i.e. the smallest proportion of ‘not sures’) was for their elected representative.

Was there anywhere that the two warring sides, both alike in dignity, found common ground? There was one case of star-crossed compromise: AI-phobes and -philes overwhelmingly thought AI-written Saint articles were a bad idea. Phew. 

What overarching conclusions can you draw from this whole exercise? Perhaps that AI-philes are the kind of people who will seriously think about DMing their mother AI generated slop on her birthday; perhaps that AI-phobes feel more strongly about their MP’s essay-writing skills than their plane flyer’s plane-flying skills. 

I asked an academic from another university her opinion on the survey’s results. “That’s very depressing,” she told me. “I just find the thought of [AI] so overwhelming that I don't know where to begin…I prefer not to think about it.” She teaches an essay-based subject; she doesn’t know how AI will affect the humanities long-term. “I imagine that students think they’re getting away with something…[but] they’re hurting themselves.” 

What do you make of all this, dear reader? Are you left fuming? Smugly nodding? Yawning? Maybe you don't know your opinion yet. That's alright — ask your friends. Think on it. Or, here’s an idea: open a new tab. Let ChatGPT do the thinking for you.

Illustrations: Grace Robinson