That may well be true for now, but it's open to the argument that it's only a matter of time, and that improvements in the technology will produce things that are richer, stranger.

That in turn may or may not be true, but in any case it doesn't matter. There's a fundamental and insurmountable reason why AI can't do art, which is about the nature of art itself.

Two groups of people are promoting AI, with some overlap between them. The technologists are interested in refining the process. The business people are interested in the product, and what it will enable them to sell.

Both groups see "content" as product, as something to be consumed. It's essentially a one way transaction: the product gets pushed out there, and the consumer enjoys the experience it delivers.

In this view a pleasant tune is a pleasant tune, a good read is a good read. It doesn't really matter where it comes from.

That's tenable if you think creative work is just a form of entertainment, but art by definition is more than entertainment. It's something capable of stimulating an experience in us that goes deeper than or beyond entertainment. I don't want to be pompous about this. It's just a matter of definition. For sure there are things that may emerge from the world of entertainment to become something more, something more profound, but that’s the point: they’ve crossed a boundary from one world to another.

And in this deeper experience, the notion of authorship matters.

It's not a straightforward notion. We may know little or nothing about a writer (I'm going to talk mostly about writing, though these ideas apply to other art work too), but even when that's the case, we will begin to imagine the personality behind the work, because part of what we seek in our response (and part of what the creator seeks too in the creation) is an imagined transaction that may offer the gratification of proving true. In trying to understand what's going on, we're making connections, between ourselves and another mind.

I'm not suggesting that this is all art does, only that it's a fundamental part of what art does.

(I'm well aware than in the later part of the last century a substantial critical movement wanted to view a creative work as a social artifact, minimising the significance of this imagined transaction between minds. Although this movement might have given us occasionally interesting insights, it always seemed to me to miss the point, ironically telling us more about the critic than it did about the work.)

Another way of gauging this is to consider how we feel when we discover that an artist didn't share our values. This doesn't mean we can easily dismiss the work, or fail to value it, but knowing that TS Eliot was antisemitic does matter. Where it crept into his poetry we can fairly be angry and dismissive: the real personality matters. We can still admire Four Quartets, for its other virtues, but the admiration must be tempered by what else we know, even when the verse is free of apparent racism.

What we know matters to what we can imagine, and what we can imagine defines our response.

Equally while I admired Johnny Marr’s guitar playing, I never liked Morrissey's songs, and having him reveal himself as a true dickhead offered the easy pleasure of corroboration.

In this light we can imagine being confronted with an anonymous poem, or story, looking for the good in it, and finding it. But it makes a legitimate difference if we then find out it's been written by an AI. It’s not simply that we’ve imagined something that wasn’t there: we can legitimately feel that our trust has been betrayed.

That's because the work doesn't stand as a self-sufficient artifact, a limited bundle of connected words.

Creative writing is not just about the grammar or vocabulary of the sentences. It's a social interchange between author and reader. Take the author away and you've fundamentally changed the meaning. In a sense you've made it meaningless.

AI, at least as built on LLMs (large language models) is nothing like a mind, and so is categorically incapable of producing art, with its work of connection. Whether future AIs could develop consciousness, something like mind, is beyond my scope here, but on current evidence I have serious doubts.

Not all writing aspires to the condition of art. I've made my living as a business writer, a living currently threatened by AI, because business writing isn't art, though it can be creative. Could the technology replace me? Unfortunately the answer is yes, up to a point, but the point is the interesting thing.

For the last few decades business writing has been dominated by brand thinking, and brand thinking revolves around the idea of personality, of the corporation having an identity we might begin to relate to in much the same way we might relate to a friend.

The difficulty here is that the notion of corporate identity is in truth a fiction. Brand people will tell you the identity needs to be rooted in a true personality, but this is to mistake consistency for personality, when real personality is not a corporate trait (how could it be, when any corporation is a mass of participating identities?)

Only people have personality, and even our personalities are subject to considerable inconsistency, because inconsistency doesn't matter when confronted with the unity of a human being. We are pushed instead to talk of a complex or contradictory personality, but that's not the stuff of which brand identities are made. A corporation doesn't have that intrinsic unity. It's a fiction that the idea of brand tries to construct, and impose. The result, ironically in this context, is that brand identities can easily seem more robotic than human.

AIs are extremely well-placed to deliver consistency in communication. But in the end as humans we seek connections with other humans. If a business seems robotic in the way it speaks, we will naturally and quite properly switch off. I don't doubt that "be less robotic" can be part of the brief to an AI, but you have to wonder whether it might be more rational, and probably successful, to ask a human to do the job you're trying to make a robot imitate.

Maybe we can get to the point where we understand that engaging people is not a matter of following the rules and parameters you have given yourself, but of being able to imagine what your reader might or might not be thinking, and then surprising them into attention.

For that you will continue to need people like me.

If you’ve enjoyed this piece take a look at a companion essay on why we write, and should carry on doing so