Writing a play about artificial intelligence in 2026 is a bold move. We've entered an era of AI fatigue: anyone who is not obsessed with AI is now sick of how much discourse there is around it. People are turning off the news right as the warnings in the media become increasingly existential. We've entered the stage of denial, of wilful blindness, of don't-remind-me-that-AI-will-wipe-out-humanity-I-don't-want-to-talk-about-it. 

Who, in this context, would spend an evening watching a play about AI? Would it be possible, right now, to write an AI-related play that doesn't immediately turn people away? To cut through the noise, the grey sludge, of AI discourse?

Tom Holloway’s latest play Eliza is a solid effort. In an era of perennial newness—where AI develops so fast that novelty has become the new normalcy—Holloway’s point of difference lies in his pivot towards history. Eliza is set in the 1960s, depicting the birth of an artificial intelligence in the office of Dr Joseph Weizenbaum (Dan Spielman), aided by his secretary Becky (Manali Datar) and research assistant Ron (Hamish Michael). A joke made by Betty about her therapist’s repetitive, ‘neutral’ form of questioning, which consistently turns the attention back to the patient, prompts Weizenbaum to wonder if a computer could be trained to ask similar questions.

In many ways, Weizenbaum is a direct reincarnation of Victor Frankenstein: an ambitious, intelligent social recluse who creates life and then regrets it, haunted by the capabilities of his own creation and how quickly it slips out of his grasp. He even has Frankenstein’s troubled, trauma-filled German childhood driving his quest for exceptionalism.

However, what distinguishes Weizenbaum from the trope is that he doesn’t hold lofty ideas of his own genius; he doesn’t believe that he has the capacity to create life. Rather than setting out to create artificial intelligence, he hopes to prove that a computer can convincingly pass as human, but still not be intelligent. He wants to mock the Turing Test instead of proving it.

Aside from this quirk, though, his character feels underdeveloped. I found myself wanting a clearer link between his childhood experiences and adult persona. Some of the most intriguing parts of the play are Weizenbaum’s disclosures of his unnerving, highly specific childhood traumas. But it seems as if Holloway uses these for shock factor alone instead of considering how Weizenbaum's character might be shaped by them. The adult Weizenbaum reveals his traumas but his character doesn’t bear their imprint, and comes off as somewhat two-dimensional as a result.

Similarly, the production’s aesthetic seems torn between three directions. For most of its duration the play embraces a mundane and distinctly 60s office aesthetic; it’s beige, prosaic and retro. Then there are moments of slick technological futurism, where Weizenbaum stands alone, bathed in blue light, engaging in a therapy dialogue with a voice-over Eliza, centred amid symmetrically arranged towers of files. Finally, there’s a brief foray into Kafka-esque surrealism, where Weizenbaum starts spitting out bird feathers, screams into a box, and wraps the office clock in his jacket. Holloway’s use of contrasting aesthetics is daring but disbalanced; selecting two rather than three, and creating a more equal interplay between them, might have given the production more cohesion and intentionality.

It feels as if Eliza is caught between experimentation and convention. But it’s when the play leans into the familiar that it’s at its most haunting. When Weizenbaum and Becky first begin to imagine the kinds of prompts a computer could be trained to recite —‘tell me more’, ‘go on’, ‘why do you feel depressed?’—the eeriness stems from the fact that we know how this story ends. We’re already at the ending. What’s disturbing about Eliza is that we’re seeing this narrative of heedless AI genesis play out right now, in the mid 2020s; for Holloway to reveal that it has already taken place, sixty years ago, is to show us that humans haven’t learnt. 

As we approach the play’s conclusion, though, Holloway offers us a way out, a chance to learn. In a session with his (human) therapist, Weizenbaum wonders if technology and artificial intelligence appeal to humans because we’ve become fixated on knowing answers: when, in contrast, what actually makes us human is our capacity to sit with questions. To hold onto wonder. I found myself wishing this conception of humanity had been more central to the play—I wanted Weizenbaum to expand on this, to keep talking about it. But maybe that’s just because I’ve become too dependent on answers. Maybe I need to just sit with my questions. Holloway’s Eliza feels incomplete, but the play itself disputes the importance of completeness. Weizenbaum leaves us with the task of extracting sustenance from the absence of answers; to revel in the existence of questions that can’t be answered. Eliza is thought-provoking, daring and imperfect, but perhaps that’s what human life is meant to be.