Reports of English literature’s death have been greatly exaggerated

Strangely, a chronic problem has become acute with the advent of artificial intelligence. Andrej Karpathy, one of the founding members of OpenAI, caused a bit of flutter earlier this year when he tweeted: “The hottest new programming language is English.” What he meant is that increasingly you don’t need to be able to code to code.

A friend of mine who works in the tech industry points out that the deep learning algorithms and transformer models created by the likes of Google, Meta, and OpenAI among others in the past few years didn’t create Large Number Models; they created Large Language Models (LLMs).

Anyone can therefore write prompts for ChatGPT and their ilk. “Natural” language is therefore becoming the “user interface” for artificial intelligence. (That sentence alone illustrates why Silicon Valley might need some help with the transition.)

As my friend put it: the invention of the iPhone put a supercomputer in everyone’s pocket, but LLMs give all of us the ability to program it. In many ways this is great news. It means that technology is becoming more democratised and accessible. It opens up a host of opportunities for those who are skilled in the use of language. The problem is, that’s not many recent graduates.

Anjney Midha, who is on the board of several AI companies, says he often sees very bright Gen Z kids struggling to write clear prompts because they mostly communicate through broken or pidgin English: “Translating thought to language is insanely hard for them.”

Ethan Mollick, a professor studying AI at Wharton business school at the University of Pennsylvania, says this means that – in an inversion of the old order of things – experienced managers are becoming better coders than bright young things fresh out of university.

Nor is the problem confined to the world of tech. Universities are finding that many students arrive having never read a whole book from cover to cover. This is leading to a massive deficit in old-fashioned skills that turn out not to be so old-fashioned after all.

For the past few years, Kingston University has been asking businesses what skills they need but currently aren’t finding in potential employees. Top of the list are the ability to communicate, analyse, adapt, problem-solve and think creatively.

What’s more, it turns out that computers can learn to code far quicker than humans can. They can easily be taught how to ace exams in maths and science. But even the most sophisticated generative AI struggles with English literature papers. There’s a clue here.

In a recent interview about AI, the mathematician Terence Tao said: “I think at the frontier, we will always need humans and AI. They have complementary strengths.” So, contrary to the prevailing doom-mongering about the relentless rise of AI being about to damn the humanities to perpetual irrelevance, might the very opposite be true?

With such a large supply and demand mismatch, you’d assume the market will eventually correct itself, but perhaps things can be helped on their way. Colin Hughes, the head of the country’s largest exam board AQA, argued the GCSE English language needed to be rethought because it was “not very inspiring” and “a bit too mechanistic”.

One obvious way to update the syllabus would be to teach the writing of clear, succinct and unambiguous prompts for artificial intelligence chatbots. English literature could also be made more relevant.

That is not – repeat, not – about bemoaning the canon for being too “male, pale and stale”, as Sharon Hague, the managing director of Pearson, recently did. Nor should we point the finger at “wokeness” for killing off the English degree.

Such tensions have always existed in the discipline. So academics can continue with their squabbles about which voices are most marginalised, but only after pointing out that studying literature is a crash course in empathy, that almost all careers require an element of storytelling and that the only way to learn how to write well is to read lots.

Harold Bloom argued that deep reading fostered higher order thinking. An education in the humanities or the liberal arts also makes students more adept at dealing with nuance and expressing opinions based on value judgments. These are useful skills for dealing with an unpredictable future and a world composed of various shades of grey.

None of this is going to result in an immediate stampede of people signing up to study English literature at university. You don’t need to understand Chaucer to write clear AI prompts. What’s more, you shouldn’t really need any better argument for studying great art other than for its own sake.

But if reading whole books and writing essays is no longer a given, then those who can will have an edge over their peers. And if more students can be persuaded that’s a good way to become more employable, a reasonable proportion of them will go on to study English literature at A-level and a reasonable proportion of them will go on to do so at university.

It’s only common sense that if you are worried about the rise of machines and robots stealing our jobs, it’s better to lean into the stuff that AI finds trickier to do. The not-so-secret ingredient is right there on the packaging; they’re called humanities for a reason.