Is it fair to have one set of rules about AI use for academic staff and another for university students?

Is it OK to use AI? Well that all depends. I think the answer has two levels. One level is universal – with factors that apply to everyone, whoever they are and whatever they are doing. The other is individual - with factors that depend on who is using it, and why.
I was prompted to write this post because I keep hearing a particular criticism of universities, namely that staff discourage and sometimes attempt to punish students using AI in their submitted assignments while happily using it themselves. On first sight that looks hypocritical, but a different set of rules for staff and students is not, I think, a double standard. It is the difference between equality and equity.
The Universal Considerations
There are at least two (probably more) factors that everyone who is using AI needs to think about - whoever they are and whatever they are using it for.
The first is the environmental impact of GenAI data centres. A single text prompt is relatively cheap. Google puts its median Gemini prompt at 0.24 watt-hours and about five drops of water - but the International Energy Agency expects data centre electricity use to more than double by 2030, with AI the main driver. A professor's prompt costs the environment the same as a student's.
The second is what AI can do to our thinking abilities. Offloading memory or arithmetic to a tool is nothing new, but researchers at Wharton have described a further step they call 'cognitive surrender'. That means adopting an AI's answer with little scrutiny, so that it simply becomes yours. Academics are not immune either. A survey of 319 knowledge workers found that the more people trusted generative AI, the less critical thinking they applied.
These shared costs are some of the reasons for everyone who chooses to use AI to use it with purpose. But they don't help us decide whether a particular use is OK.
The Individual Considerations
You will likely be familiar with the main points of this debate. Students review their lecturers' AI use on sites like Rate My Professors, and scour slides for words ChatGPT is fond of, such as 'delve'. Lecturers, meanwhile, take to social media to complain about AI-written essays and the existential crisis AI poses to Higher Education credentialling. Essentially I can see four points that keep coming up.
'Students should be allowed to use AI'
It's just the latest tool. Calculators met a similar backlash in the 1970s, as did spell-check and Wikipedia later. Universities should redesign assignments, not rely on vague policies and flawed detection tools.
Professionals use it, so why can't we? Students should be able to use AI for everything because that's what happens now in the workplace.
We need the skills. Good AI skills are essential for employability, yet students are not getting the support from their institutions they need to develop them.
'Students shouldn't be allowed to use AI'
It short-circuits learning. AI is a shortcut that undermines the processes that lead to learning, so some academics mark down when AI misuse is suspected.
The public largely agrees. In a TES survey of 2,221 adults, almost nine in ten disagreed with students using AI to complete school coursework entirely.
'Academic staff should be allowed to use AI'
It helps them do the job. Generative AI makes Academic staff better at their jobs, and acts as an automated teaching assistant amid heavy workloads.
Students want more comprehensive feedback. Students want more consistency, speed and volume of feedback.
It frees up time which can be put to better use. AI use enables academic staff to give students responses instantly, and contact time can be spent in deeper discussion.
'Academic staff shouldn't be allowed to use AI'
It's hypocritical. A business student found her professor had used ChatGPT to produce course materials, despite a syllabus barring students from unauthorised AI use. Imperial's student newspaper found student AI use strictly regulated, but no firm guidance on how much lecturers could use it.
We're paying for humans. Students pay a great deal, to be taught by people, not by an algorithm they could consult for free.
Feedback will get worse. AI marking will homogenise feedback, so students increasingly hear the same things. AI is not yet good enough to mark reliably, rewarding style over substance.
Academics will lose something. Outsourcing feedback is a false economy, because academics lose the craft of feedback and the relationships it builds. Keeping a human in the loop is harder than it sounds, as academics are swayed by AI feedback instead of reviewing it critically.
Then who is learning? 'If both the students and the teachers are using it, then like, who's learning?'
The strongest points on every side are about purpose. The worry for students is that AI replaces their learning. The worry for staff is that AI replaces the human judgement and relationship that are fundamental to education.
Equality is not the same as equity

You know the cartoon where three people of different heights are trying to watch a ball game over a fence. In the 'equality' panel, each stands on an identical box, and the shortest still cannot see. In the 'equity' panel, the boxes are shared out according to need, and everyone can watch. It began as a PowerPoint sketch by Craig Froehle in 2012 and was later redrawn by Angus Maguire for the Interaction Institute for Social Change.
If we wanted to treat everyone the same with AI, we could give everyone the same size step - regardless of whether it was high enough to see over the wall, or whether they needed it at all. In contract, equity asks what does each person need, given why they are here?
Staff and students are not at university for the same reasons.
One of the main reasons students are at university is to learn. The effort of working something out - drafting, getting stuck, being wrong, trying again - is not an obstacle standing between them and their education. To a large extent it is their education. When a student uses AI to get around the work, they are doing themselves out of what they came for.
Staff are at university for different reasons. One of the main reasons is to help students learn. When an academic uses AI to help produce better feedback, the person who benefits is the student.
So a different rule for staff and students is not hypocrisy. It is treating people according to their needs and their purpose. I think it is OK to have one set of rules for one and another set of rules for the other.
I am not recommending unfettered, uncritical or hidden use of AI by academics, or using it in place of human judgement or criticality. The critics of academic staff AI (mis) use make fair points. But I think they are describing bad practice, not a reason for no practice at all.
I am not arguing that 'academic staff can do what they like'. I am arguing that AI used openly, critically and in the service of students' learning can make academics more effective in their role - and that this is a different thing altogether from a student using AI to bypass learning.
So, is it fair to let academic staff use AI and not students?
It depends. Some of the costs fall on everyone, and universities should be as honest with staff as with students about them. Beyond the shared factors, though, the question 'should I use AI here?' depends on who you are and what you are trying to do.
For a student, the test is whether it helps them learn or helps them avoid learning. For an academic, it is whether it makes them better at helping students learn.
These are different tests because they are different jobs. Perhaps what universities could do better is explain that reasoning, so that students hear a principled position rather than what sounds like a double standard.
This post was developed with the use of GenAI.



Comments