Daily Maverick
Dailymaverick logo

Opinionistas

This is an opinion piece. The views expressed are not that of Daily Maverick.....

The fundamental clash between artificial intelligence and teaching the humanities

Clive Glaser
By Clive Glaser
10 Apr 2025
Facebook
5
AI is ultimately anti-humanities. It is extremely broad, yet shallow. It is anti-reading, anti-writing, anti-complexity. It does not encourage individuals to think critically or write creatively.

I teach history at a university and, like other colleagues in the humanities, I have had to do some serious rethinking in the face of the artificial intelligence onslaught. I am convinced that AI can be a very valuable tool in the sciences, even in teaching science. But I am struggling to see how AI brings any value at all to teaching in the humanities.

Much of the discussion to date has been around the scourge of cheating. How do we prevent students from simply generating answers using AI tools? This is an important discussion.

But for me, the problem runs much deeper than cheating. It reaches to the very core of what we do in the humanities.

So, let’s start by asking what we as teachers would like to achieve in the humanities. Of course, we would like students to absorb a certain amount of detail about human history, ideas and behaviour.

But, far more importantly, we want to impart skills. Students who do humanities degrees should learn how to read extensively, effectively and critically. Ideally, we should instil a love of reading and a sense of curiosity.

Emerging out of this, students should learn some writing skills. They should be able to write with clarity and structure an argument. They should be able to convey the complexity of social interaction. Dare I say it, some should even be able to write with a certain flair. We are trying to nurture the next generation of critical thinkers, artists, teachers, social researchers.

How does AI help in this endeavour?


Large language models (LLMs) are excellent at aggregating and summarising. They write grammatically and clearly. They structure information systematically. If you ask AI tools the right questions, they can tell you about internal debates and disagreements within the available digitised literature. They might even judge which arguments are weightier.

As the months go by, AI tools seem to be making fewer factual errors, especially if they are challenged with good follow-up questions. There is no doubt that cross-language translation can be enormously valuable for humanities research.

But can any of this help to impart core humanities skills? Aside perhaps from assisting with the absorption of some basic factual content, I don’t believe so.

If students turn to AI, how do we help them to learn to read, let alone read critically, and develop writing skills? AI enthusiasts tell us we should work with AI. We should, for example, let students critique AI-generated answers. I am unhappy with this. First, AI is quite capable of critiquing its own answers. More importantly, why should I let an LLM shape my teaching programme?

AI is ultimately anti-humanities. It is extremely broad, yet shallow. It is anti-reading, anti-writing, anti-complexity. It does not encourage individuals to think critically or write creatively.

So how do we work with it? Linking back to the cheating discussion, my instinct is to go back to asking students to write (or give oral feedback) in an environment completely cut off from technology.

This is not merely an anti-cheating measure. It is about pedagogy. If I assign readings to students, they are far more likely to read them, absorb them and engage critically in a space cut off from technology. They will, with guidance and careful grading, sharpen their writing skills over time.

Of course, it is far from ideal. There were good reasons we moved away from the emphasis on closed examination settings in the first place. It is desperately sad that we are forced to abandon take-home assignments; a few students really flourish through self-reading and writing. As things stand, though, I can’t see a viable alternative.

Let me stress that I am talking largely about undergraduate teaching. Many of these arguments do apply to Honours and Masters’ coursework as well (especially if the classes are large and undergraduate training has been uneven), but usually at postgraduate level we can work closely with students individually or in small classes to address these problems.

Redundant skills?


Are the skills of critical reading, creative thinking and writing simply becoming redundant? Some colleagues have argued that if we have AI, we shouldn’t fight against it. Instead of teaching these old-fashioned skills we should rather be helping students to work effectively with AI.

After all, AI can read for us, summarise, write nicely. Humans are moving into a new age. Just as we no longer need to learn how to hunt and make fires without matches, in the future we won’t have to read and write and engage critically with original texts.

I think this is a very dangerous argument. Even the most gung-ho advocates of AI acknowledge that AI tools are only as good as the input material. So, who, then, is going to feed in quality research and writing? How do we advance knowledge? What will happen to art?

If we lose our fundamental humanities skills, AI will eventually operate on an ever-expanding self-referential loop: balanced, wide-ranging at best, but bland, passionless, unoriginal.

It goes without saying that we have to learn how to work with AI. But AI has very serious limitations. And we can’t allow a passive dependency to set in.

There are vital human skills which need to be passed on to the next generation. DM

Categories:

Tags: