What ChatGPT can never do

It’s a belated summer here in Belfast, where I’m attending the UACES conference. As well as a lot else, we’re running a series of L&T panels, including one on assessment.

The prompt for this was the bruhaha over ChatGPT and Why All Assessment Is Impossible (I paraphrase): while we might understand that the problems lie much more with the assessment design than the scary AI, that’s still not really filtered through to all of our colleagues.

Our discussion was really stimulating, both for its breadth and for its reflection on what we are trying to achieve in our assessment – and, by extension, in our teaching.

A concern that was raised was that if AI can produce more encompassing integration of knowledge than any human, almost instantaneously, does this mean the ‘end of thinking’ for our students? Even if the lurid framing might rile, the idea is not to be dismissed.

For me, this prompted the thought that even in an era of all-knowing technology there is still a clear role for us as teaching, instructors and facilitators: to help emancipate our students by giving them the tools to build their agency in the world.

We frequently talk about building self-reflection and criticality in our students. In an age when the challenge is parsing and navigating through too much information, getting students to make informed choices about what to use and how to use it is essential.

And we do that because it gives them a way of standing more firmly in the world, to achieve what they want to achieve.

We don’t (I think) want to produce reproductions of ourselves, but autonomous individuals who can both define their own purpose in life and find ways to realise that purpose.

In this framing of education-as-emancipation, it becomes irrelevant what AI can do, precisely because it’s AI doing it, not the student. As one colleague noted, almost all our students don’t come to university to cheat, but to learn and to develop themselves.

Put differently, if everyone can just turn to ChatGPT, then what gives you the edge is understanding that that technology can and can’t do and understanding how you can use that to your own purposes.

This points to assessment that valorises reflection and critical engagement with knowledge and with arguments, so that the student is able to apply such tools to other situations. It also suggests that we as instructors have to spend more time on assessment that is grounded in individual experience and that recognises there is as much value in being able to articulate your self-awareness as in nominal achievement of a particular task.

A case in point in our discussion was groupwork. Yes, you can mark how well a group functions together, but we know from our own lives that sometimes we have to work with people who aren’t our optimal partners [cough], so there’s as much value in understanding how to cope with and mitigate that scenario as there is in everything being super-positive.

When we can pick up a device and find out a pretty decent amount about any given subject at the drop of a hat, ‘knowing stuff’ isn’t useful in the way it used to (even if we go to lots of pub quizzes), and we need to recognise that in all of our teaching practice.

Indeed, you might argue this is a great example of being self-aware and self-critical that we should be applying to ourselves, even as we apply it to our students.