Joanna Pidcock: The Doom Squad
Doomers, by Matthew Gasda, directed by Zsuzsa Magyar. Rose Lipman Building, London. 18 September – 3 October 2025.
Not so long ago, someone I knew from college posted some images on Instagram. They were AI-generated paintings of his house, rendered in the style of Van Gogh, Monet, Rembrandt and Dali. Crazy time to be alive, he wrote, feels like the end of times for Artists. It feels almost redundant to say that the paintings were very bad – the Dali particularly egregious – but I will, and they were. The past year has been a tipping point for AI, in terms of both widespread uptake, and handwringing defeatist discourse about the same, brilliantly explored in this recent n+1 piece which describes AI generated material as “itself a waste product: flimsy, shoddy, disposable, a single-use plastic of the mind”. AI image generation feels like yet another nefarious frontier that is being crossed with abandon, my college acquaintance one of many millions of people blinded by the novelty of having a frivolous aesthetic demand immediately granted, like a low-rent, latter-day Medici.
Sitting in the slop heap at the end of this grimly momentous year, the question – can AI make art? – feels redundant. Technically, sure, but can it make good art? The benchmark for what constitutes good art is utterly subjective and yet appears to be another metric becoming increasingly polarised and calcified by the current political climate. Should it make art? Should it make anything?
The question that came to me as I watched Matthew Gasda’s play Doomers in September, however, was not about whether or not AI can ever make good art, but rather: can good art ever be made about AI?
*
There was a lot about the experience of being at Doomers that had the kind of uncanny hollowness that comes with reading or seeing something that has clearly been AI generated. It was quite difficult to tell who produced the work and how – comparing the size of the team with the fringe scale of the production, someone wasn’t getting paid. The play was in two halves, but each half had a completely different cast, with no crossover between the two. Most bafflingly, the situation at the beginning of the play – a Sam Altman-esque CEO has been fired from the AI company he created following dangerous and worrying developments in the technology, and his team and the board are deciding what to do next – is exactly the same at the end of the play.
That is to say: across a two-hour runtime, there was no change whatsoever in the dramatic circumstances. The play comprises two large conversations about AI, with two completely different groups of people, across two hour-long halves, where nobody’s opinions or conditions materially change. No real stakes, no narrative movement, just tech chat.
At one point, I scribbled on my programme: is this even interesting? Theatrically, the answer is a resounding no. Perhaps it was the subject matter, perhaps it was the use of AI in the writing process (Gasda has mentioned in interviews that he used Chat GPT and Claude while writing the script, especially for the more technical conversations); either way, it felt as though someone had asked a Large Language Model to write an argument between five people who all have slightly different opinions on AI, twice, and then that conversation was staged with all the theatrical flair and interest of an actual corporate board meeting. As one of the characters said, “this conversation has like, devolved”. Generative AI, and LLMs in particular, are designed to make the most obviously rational choice in every situation, weighing up the likelihood of one word following another on a statistical basis. Similarly, the play and its production seemed to consistently make the most obvious choices – there was not a single moment where I was surprised. This was bland, smooth, unscented theatre.
As someone who makes theatre, I find it darkly funny that such a precarious and unreliable industry is also, largely, AI-proof. While AI can access every play in the public domain, tell you about the history of theatre, find images and recordings of productions, and possibly even call up the technical specifications of every theatre on earth, it fundamentally cannot understand what a play actually is. All of these data points are the context and structure and noise around a play, which has to be a live experience, a lived experience, one from which AI by its nature is necessarily excluded.
In perhaps his best play, Arcadia, the late, great Tom Stoppard writes about chaos theory and the mathematics around “the deterministic universe” – the idea that “the future is all programmed like a computer […] because everything including us is just a lot of atoms bouncing off each other like billiard balls”. However, one of the characters argues that the reason why the maths doesn’t work out is the unpredictability and liveness of sex and desire, “the attraction that Newton left out”. In another sense, an algorithm could have every single data point about a piece of theatre, every word spoken, every sound and lighting cue programmed, and still never be able to process what actually makes it work – the complication in the live moment where a caught breath on stage suddenly makes every person lean forward in their seats, an irritating cough in the audience that pulls them all back again.
People who work with or in AI seem enthralled by it, and in conversation are often breathlessly delighted by its possibilities and awed by its power. As such an assumption is often made that other people will also find it interesting. Perhaps that was the greatest sin of Doomers: to assume that the subject matter itself was interesting enough without having to make the scenario, or the relationships, or the theatrical language itself interesting too.
*
An argument around art and AI is the fact that human artists too are operating on a kind of Large Language Model, reframing and regurgitating their own influences endlessly in order to produce work. Everything has been seen before and there is nothing new under the sun. This jaded and pessimistic view of human intellectual and artistic endeavour aside, it is true that some artists do, in fact, do this, and the resulting work often feels flat and derivative. It is also true that most artists make work that is in some way a response to other work, in form and content. What this argument misses, however, is that what artists are generally responding to, when they make work that is influenced by someone or something else, is not so much the work itself, but how it made them feel. The generative impulse is not the data points of composition and colour, syntax and structure, but rather the live moment of encountering the work and the way that it moves through the body’s avenues of sense and perception and somehow, miraculously, often nonsensically, shakes something loose in your soul.
Perhaps I’m being romantic and naïve, a luddite wilfully blind to the fact that the AI age has well and truly arrived and isn’t going anywhere. Perhaps, as the title of the play suggests, I myself am a “doomer”. But by outsourcing our thinking to an endlessly mimetic algorithm that can only harvest information and repeat it, we are sacrificing surprise, and I find this existentially deadening. A world without the delight and terror of something genuinely unexpected, the vertiginous elation of an original thought, of something completely new – no amount of convenience and instant gratification could possibly justify this.
To be in the world at the moment is to be subject to a ceaseless deluge of information and input, much of it artificially generated and possessing a kind of smoothly uncanny frictionlessness. There is a deadening effect to consuming so much uninventive and unsurprising material, so little of which bears any resemblance to the grit or effort of actually being alive. It is not too much to ask that the art that I meet move me in some way, or shake off the oily almost-realness of everything else. For it to be a perfectly ripe, outrageously alive Mediterranean tomato, with an aroma that seems to exist in three dimensions, in a world of Huel meal-replacement beverages.
Theatre is an effortful artform. It defies the logic of the market, it is impractical and expensive, and requires many different skilled people to create. In the making and the sharing, it is completely engaged with all the irritation and challenge and shocking joy of human interaction. It is a collaboration that is often full of friction, but friction generates energy and impetus. A group of people start with an empty room and build an entire world. People saying that the days of artists are over underestimate our need to be genuinely engaged and interested by something. Man cannot live on Huel alone.
AI is here, but using it, especially when it comes to making art, is a choice. Our breathless engagement with it only feeds it and allows it to refine itself – we are giving our intellectual labour to the machine so that it can continue to take it away from us. We are allowing ourselves to be convinced that we can’t, or shouldn’t, engage in critical thinking within and about the world. In using it, we are smoothing and planing our own brains in order to better receive the denuded, unoriginal, easily digestible slop that it will produce for us.
Can good art ever be made about AI? Perhaps, but it likely won’t be about the AI technology itself, or the people building it, at least for as long as they see themselves as the masters of the universe and their creation as singularly, unquestioningly compelling. Perhaps it’s too soon for the first truly interesting piece of work about AI – it’s certainly not Doomers. On the night I saw the play, the audience watched with warm interest, but the person laughing the loudest throughout, who seemed most engaged with the work on a human level, was the playwright himself, sitting in the front row.
Joanna Pidcock is an Australian-born, London-based theatre director, dramaturg, and writer. She has worked at theatres including the National Theatre, Shakespeare's Globe, L'Odéon, die Schaubühne, the Soho Theatre, and New Diorama, among others. As a prose writer, she was shortlisted for the 2021 Fitzcarraldo Editions Essay Prize, and her work has been published in TOLKA Magazine.
