We shouldn't fear ChatGPT in education — we need to work with it

There is no point pretending ChatGPT doesn’t exist, and trying to prevent its use by implementing honour codes and harsher plagiarism penalties will not work either

Instead, by finding ways for students to work with ChatGPT, we can reduce the likelihood of cheating, make the assignments more interesting, and really develop the critical analytic abilities that we were seeking all along.

In recent years, artificial intelligence has been making great strides in various fields, including education. 

One of the most remarkable examples of this is ChatGPT, a language model developed by OpenAI. 

While ChatGPT's primary purpose is to assist users in generating human-like text, it has also made significant contributions to the field of philosophy. 

This, in turn, has had a significant impact on educational assessment, enabling educators to evaluate students' critical thinking skills in new and exciting ways.

… or so the enthusiasts say. 

The doomsayers, by contrast, are sceptical. Here at UCC, for example, I’ve heard more than a few colleagues echoing Socrates, who, in Plato’s dialogue Phaedrus (370 BC), expresses similar worries about the invention of … wait for it … writing. He says:

"This invention will produce forgetfulness in the minds of those who learn to use it ... you offer your pupils the appearance of wisdom, not true wisdom, for they will read many things without instruction and will therefore seem to know many things, when they are for the most part ignorant and hard to get along with since they are not wise, but only appear wise."

Fast forward a couple of millennia and the worry is the same; students might use this new technology to write their essays for them, in ways that escape the notice of standard plagiarism-detection tools, giving rise to a new form of turbo-charged AI-powered cheating.

What’s particularly unsettling is that language use has long been regarded as the hallmark of intelligence. 

The philosopher René Descartes (who you may remember from such lines as “I think, therefore I am”) held that linguistic ability is what distinguishes humans from animals; he thought it was evidence that we have immortal souls, whilst they are mere automata.

The Turing Test

Despite his general scepticism about AI, Descartes thereby anticipated another famous 20th-century development in the field: the Turing Test. 

In his landmark 1950 essay entitled Computing Machinery and Intelligence, Alan Turing suggested that instead of asking the somewhat nebulous question “Can a machine think?" we’d be much better off addressing the more concrete issue: “Could a machine deceive a human interlocutor into thinking that they were conversing with another human?”

Turing’s idea was that we can never really be certain what, or that, another person (or machine) is actually thinking, but that conversational ability is like the tip of an iceberg: a pretty good and generally reliable way of inferring that there’s a whole lot more going on beneath the evidence at the surface.

I sometimes joke with my students that the essays and exams that I ask them to write serve much the same purpose as the Turing Test.

I can’t really be certain that they’ve learnt and understood everything that we’ve covered, but the 2,500 words that they hand in at the end of the semester is like the tip of an iceberg; if I construct the assignment carefully, it gives me a pretty good and generally reliable way of inferring that they’ve learnt a whole lot more, as a backdrop to the evidence presented in the essay. 

So the worry I mentioned earlier is that since ChatGPT can clearly pass the Turing Test, then it could also pass university courses — on behalf of cheating students — in the same way.

… Or could it? Notice that I said “...if I construct the assignment carefully.” Many “standard” paper topics (e.g., “Describe and explain theory X about topic Y”) are the sorts of prompts for which ChatGPT can produce a perfectly serviceable passing essay in response.

But these have never been good essay questions, because they encourage exactly what Socrates was worried about: rote learning and boring, formulaic answers.

Reduce likelihood of cheating

Instead, by finding ways for students to work with ChatGPT, we can reduce the likelihood of cheating, make the assignments more interesting, and really develop the critical analytic abilities that we were seeking all along.

I imagine that this is the reasoning behind the recent announcement that pupils will be allowed to use ChatGPT as part of their coursework for the International Baccalaureate.

So rather than playing the role of ostrich or cop (and either pretending ChatGPT doesn’t exist, or else trying to prevent its use by implementing honour codes and harsher plagiarism penalties), this semester I plan to “lean in”.

In one of my classes, students must first ask ChatGPT to answer as if it were René Descartes (it refuses to impersonate someone, but it responds well if you start with “Let’s play a game...”) 

They must then conduct an interview about its views on the nature of the mind, using ChatGPT to role-play a dialogue, and then critique the answers based on what they know from the philosophical texts.

In another class, I’ve asked students to conduct an actual Turing Test with ChatGPT, and then to evaluate its performance based both on what they’ve read in Turing’s paper, and on what they know about how ChatGPT works.

In class, we have discussed how ChatGPT works, its strengths and its limitations, with a view to enabling students to treat it just like they would any other source.

Pocket calculators and spell-checkers didn’t imperil the maths and English skills of previous generations.

We’ve learnt to use the written word quite well, despite Socrates’ worries.

So I believe we can do the same thing with new AI technologies too. 

For that reason, on balance and contrary to the doomsayers, I am inclined to side with the slightly more optimistic opening paragraph above, even though it’s a sentiment for which, I’m willing to admit, I had a “co-author”:

User: Write the opening paragraph of a short article for the Irish Examiner, about ChatGPT, philosophy, and its impact on educational assessment.

ChatGPT: In recent years, artificial intelligence has been making great strides in various fields, including education. One of the most remarkable examples of this is ChatGPT, a language model developed by OpenAI. While ChatGPT's primary purpose is to assist users in generating human-like text, it has also made significant contributions to the field of philosophy. This, in turn, has had a significant impact on educational assessment, enabling educators to evaluate students' critical thinking skills in new and exciting ways.