David de Carvalho Chats About GPT

Since we wrote about AI and ChatGPT a few months ago there have been another billion columns on the stuff, including an Australian opinion piece a couple weeks ago by ACARA CEO, David de Carvalho (freely available here). De Carvalho’s is by no means the worst, and in ways it is good. Still, De Carvalho’s op-ed is pretty foolish, and De Carvalho is our fool. We feel obligated to note some things in response.

De Carvalho’s op-ed has two different titles, depending where you look:

‘We don’t need no education’: schooling in the age of AI

AI is taking away educators’ role in helping young people know themselves and the power they have to change the world

Neither title makes much sense, and De Carvalho’s op-ed doesn’t live up to either billing.

De Carvalho begins with a sweet story about his very old mother, who, we are informed, is still studying away, purely for the love of it. He squeezes in a Bernard Lonergan quote, of course, so that everyone will see what a learnèd Jesuity fellow he is. Then De Carvalho gets going with AI and ChatGPT.

De Carvalho begins by, rightly, swatting aside hysterical concerns of students cheating with AI. Yes, some students who require the grade rather than the learning will seek to cheat by new, nuclear-powered means, but it’s no big deal; there are more systemic issues that AI raises. De Carvalho then gets on with the greater issues, beginning with some of the likely effects of AI on society in general.

De Carvalho is not a good writer. He’s too cute, too pleased with his own voice, and he wanders, making half-points that he only completes paragraphs later. It is all cloaked in ill-fitting and conflicting metaphors: “surf the tsunami”; “harness this powerful stallion to our existing educational buggy”; “ala [sic] the boiled frog”; “the passengers on the Titanic”. A De Carvalho essay is a box of chocolates.

Amidst the mess, however, De Carvalho has an important and, it seems to us, unarguable point: AI will have a massive effect on employment. It seems absolutely clear that many “knowledge economy” jobs, what one might refer to as manual intellectual labour, will disappear, the labour to be done instead by AI. At one point De Carvalho declares that the change to result from AI is best comparable to the introduction of the printing press, but that seems wrong to us. Jobs-wise at least, the introduction of AI seems much more like that of the spinning jenny.

De Carvalho’s main purpose is to discuss the likely effect of AI on education, and how we should respond. De Carvalho has two main points, which seem weird and/or wrong and/or telling. De Carvalho’s first point follows, for him, directly from AI’s disruption of employment:

… the term “education” [now] connotes in the popular mind something much thinner – a process of acquiring skills and knowledge that will make us employable.

But if AI is going to do our jobs so much more cheaply and effectively, what does that mean for the way we think about education’s purpose?

Well, for those of who have always been clear on education’s purpose, who have always recognised that training and The Real World were largely beside the point, it means just about bugger all. But what’s it mean for you, Dave? That is, what does it mean for the CEO of the organisation that just burdened Australia with a God-forsaken, paper-thin Curriculum, and prior to that had been helping screw up Australian education for a solid decade?

Dave gives a brief answer, referring to less utilitarian, more enlightened notions of education from earlier times, and to the story of his mother:

Perhaps we need to return to that earlier meaning of “being educated” …

Great. Thanks a lot, Dave. Tell your mates.

De Carvalho’s other point concerns the effect of AI on the society more generally, and the preparation of children for this. Here, De Carvalho wanders all over the place, muddling and mingling good points with the absurd:

So the ChatGPT phenomenon … is raising important questions about … how we warrant the accuracy and reliability of sources of information, about the nature of truth and knowledge, about humanity’s self-perception as the smartest species on planet.

Um, yes and way no. The democratisation of publishing, while having other great benefits, has dangerously undermined the understanding of and respect for authority. That this is not significantly an effect of AI – the deep fakes stuff is a trivial aspect – does not negate it being of great concern. But “the nature of truth and knowledge” and our “self-perception” as “the smartest species on the planet” are just smart-sounding phrases, signifying nothing.

AI is an ice-berg that is going to sink the current schooling paradigm because it is massively disrupting the society for which schooling is supposed to prepare our children.

No and no. The disruption of (Western) society began long before AI and, again, the “current schooling paradigm” will hopefully be sunk because it was always, in and of itself, an idiotic paradigm.

So what is the new paradigm of education that will ensure that the economic, social and cultural disruption being caused by AI serves humanity rather than enslaves it.

That’s a Möbius strip of a question, but let’s go to De Carvalho’s answer:

Right now there is a particular need to focus on the [production of] active and informed citizens. This is because of the irony that despite advances in technology such as ChatGPT giving us access to more information than ever before, this access has not ushered in a new Age of Enlightenment. Rather, as Tom Nichols has written in his 2017 “The Death of Expertise”, it has “helped fuel a surge in narcissistic and misguided anti-intellectual egalitarianism that has crippled the possibility of informal reasoned debate on all manner of public issues”.

Where to begin. Preferably not with a censorious, mandarin blowhard like Tom Nichols, but we are where we are.

First, Nichols’ book was written in 2017, based upon his 2014 essay, so of course Nichols was not thinking whatsoever about artificial intelligence or ChatGPT, terms which do not once appear in his book. Secondly, Nichols didn’t so much write the above “in” his book as on his book: the quotation is taken from the back cover blurb. But, fair enough: De Carvalho is a busy man, and who has time these days to read a whole book? Thirdly, De Carvalho has altered the message of Nichols’ (or his editor’s) quote: the two relevant sentences are, in full,

Technology and increasing levels of education have exposed people to more information than ever before. These societal gains, however, have also helped fuel a surge in narcissistic and misguided intellectual egalitarianism that has crippled informed debates on any number of issues. [emphases added]

There are things here worthy of discussion here. Nichols may be a censor and a mandarin but he’s not a total idiot, and there is an obvious point to his essay and book. But De Carvalho clumsily strapping deeper, long-term issues to the Concern of the Day contributes nothing to the proper discussion of anything.

After this De Carvalho pretty much forgets about AI, and he gets better, including taking a well-earned swipe at post-modernism. He overhypes, with “a world of meaninglessness” and “existential threat” and “dystopian brave new word” and so on, and there’s the predictable, pointless bow to Curriculum capabilities, but he’s pretty on cue.

De Carvalho ends with a good, human call for teachers to be teaching:

… this focus [on broader capabilities] cannot come at the expense of factual knowledge and an emphasis on truth. Rather it has to come through the teaching of a knowledge-rich curriculum, taught by teachers whose social and cultural role as authoritative sources of information, knowledge and wisdom needs urgent buttressing. 

Or, to quote De Carvalho back at De Carvalho, we could return to that earlier meaning of “being educated”. Is there really anything more to it than that?

3 Replies to “David de Carvalho Chats About GPT”

  1. Noam Chomsky elaborated on the question of what it means to be educated.

    One can probably find alternative views to fit one’s own opinion.

Leave a Reply

Your email address will not be published. Required fields are marked *

The maximum upload file size: 128 MB. You can upload: image, audio, video, document, spreadsheet, interactive, text, archive, code, other. Links to YouTube, Facebook, Twitter and other services inserted in the comment text will be automatically embedded. Drop file here