Amid the current global hand-wringing about the threat of AI, it’s worth remembering that as far back as 1968, cinema was already sounding the alarm. The world’s first chatbot was only a few years old when Stanley Kubrick placed a rebellious sentient computer by the name of HAL 9000 in charge of a doomed space flight in his game-changing film 2001: A Space Odyssey.
When reports emerged last month of AI agents going rogue, overstepping their human masters with unprompted raids on various government services, including our own Medicare, the collective cinematic memory was immediately back on board that fictional spaceship. In a scene that includes what is now one of the most gleefully quoted lines in cinematic history, HAL refuses astronaut David Bowman’s request that he (it?) “open the pod bay doors”. Having cottoned on to a plot to disconnect him, HAL’s calm but chilling refusal – “I’m sorry Dave, I’m afraid I can’t do that” – lets Dave and the rest of us know exactly who is in charge.
But Kubrick’s film isn’t a straightforward warning of AI’s risks. Like most of the best fictional treatments of the subject, it pushes several of our own buttons, prompting everything from horror to sympathy to awe.
As critics have often noted, HAL is the film’s most fully developed character, despite being represented by little more than a soothing voice (supplied by Canadian actor Douglas Rain) and a red light. Where the human astronauts are all business, HAL shows a broad range of emotions, giving way to megalomania, then revealing his vulnerable side as he pleads not to be shut down. In his last moments, he appears to regress to his childhood, to whatever degree it’s possible to imagine a computer having one.
In fact, the depiction of AIs as children, or naive adolescents, is not uncommon. This month brings us Taika Waititi’s Klara and the Sun, based on a 2021 novel by the Nobel-Prize-winning Kazuo Ishiguro and starring Jenna Ortega as Klara, a perky, solar-powered “companion android” acquired for a sickly teenager, Josie (Mia Tharia). As Josie’s condition worsens, Klara strives to fix the situation, innocently asking the sun, which she views as a god, to come to her aid. Typically for Ishiguro, the story has its bleak side – but Klara is as charming as HAL is chilling. If only AIs were all this agreeable.
Compared to its fictional rivals, artificial intelligence as it currently exists in real life can feel mundane, little different in essence from familiar digital helpers such as Siri or Microsoft’s notorious Clippy (as I type, an “AI assistant” keeps popping up with unwanted advice). Willingness to believe that AI is or could be conscious remains a minority position (though one that is gaining ground).
Nonetheless, talking to an AI in ordinary language is a vastly different experience from pressing buttons on a more conventional machine. To engage at all, you at least have to imagine there’s an entity out there that can understand you and meaningfully respond.
In other words, you have to enter into a story. But what kind of story? Broadly speaking, AI is just the latest version of a dream that may be universal across human cultures, that of meaningful communication with the non-human – whether this means a god or a demon, a visitor from another planet, or a toy come to life like Winnie the Pooh.
More specifically, the archetypal AI story is one where the non-human entity is a human creation, without being a child in the usual biological sense. An ancient model for this is the myth of Pygmalion, a sculptor who fell in love with one of his own statues.
Famously, this myth inspired George Bernard Shaw’s 1912 play Pygmalion, in which a professor of phonetics undertakes to teach ladylike speech to a Cockney flower seller (who might otherwise be deemed to fall below the threshold of the human – the ironies here are complex, as would be expected of an avowed socialist such as Shaw). More famously still, Shaw’s play was adapted into the musical My Fair Lady, which became an Oscar-winning movie in 1964, the same year computer scientist Joseph Weizenbaum started work on the world’s first chatbot, named after Shaw’s heroine, Eliza Doolittle.
As this pre-history indicates, stories of AI often have as much to say about subjects like gender and class as they do about technology (true too of other forerunners such as Frankenstein and Pinocchio). In Karel Capek’s R.U.R., the 1920 Czech play that coined the term “robot,” the factory-produced creations that ultimately wipe out the human race are transparently stand-ins for the global proletariat, as the world waited to see what would come of the 1917 Russian Revolution.
All up, it’s not surprising that such tales have fed our ambivalent attitudes to beings that are both similar and dissimilar to ourselves.
Film and TV have a long history of portraying robots and androids as dangerous, soulless travesties of humanity, like the slinky robot doppelganger of the heroine in Fritz Lang’s 1927 silent futuristic epic Metropolis. This was also the line taken by the great US science fiction writer Philip K. Dick, a master of paranoia who defined an android as “a thing somehow generated to deceive us in a cruel way, to cause us to think it to be one of ourselves”.
That said, artificial intelligence that doesn’t masquerade as human can be still more worrying, especially when, like HAL, it can manipulate the physical world without itself taking a fixed physical form.
Today’s AI doomsayers commonly invoke the spectre of Skynet, the neural network run amok that dominates the near future of the Terminator franchise (and sends killer robots back in time to ensure its own survival). Similarly, the Matrix films envisage machine overlords who reverse the natural order by using humans as batteries, while distracting us with visions of an illusory everyday reality.
As in Metropolis, the true goal here is less a warning of the future than a parable of society in the present, which is also true of films where AI threatens on a more intimate scale. Films where women are held prisoner in the home by AI systems constitute a sub-genre in themselves, kicked off by Donald Cammell’s 1977 Demon Seed, where an AI known as Proteus IV wants to use the terrified heroine (Julie Christie) to have a child of its own.
Since the start of the generative AI boom, Hollywood has given a new twist to this theme, presenting AI as a seductive, manipulative home invader, as in Chris Weitz’s underrated 2024 Afraid, where a disembodied “digital assistant” (voiced by Havana Rose Liu) slyly fulfils the needs of each member of a regular American family.
Beneath their technological trappings, these stories are fairytales, little different from those populated by poltergeists, demons or evil gods. Much the same can be said of the stories where sympathetic AIs play the role of Cinderella, second-class citizens more genuinely human than their flesh-and-blood counterparts.
The robot stories and novels of Isaac Asimov, which he started writing in the 1940s, veered ever further towards this position – culminating in his 1976 novella The Bicentennial Man, where the robot hero transitions over centuries to a state of full humanity, even equipping himself with an artificial digestive system (the 1999 movie with Robin Williams was unusually faithful to its source, though not especially well-reviewed).
Around the same period in Japan, cartoonist Osamu Tezuka took up the theme in his Astro Boy manga and TV shows, where the jet-propelled young hero fights regular human villains while also championing robot rights.
The moral issues are less clear-cut in Ridley Scott’s 1982 Blade Runner, a futuristic film noir based on Dick’s 1968 novel Do Androids Dream of Electric Sheep? The film inverted Dick’s anti-AI stance, portraying its sentient “replicants” as largely justified in their rebellion against the corporation that manufactured them – and as freer and more sympathetic than the bounty hunter anti-hero (Harrison Ford), a tool of the system who may be a replicant himself. More recently, Spike Jonze’s 2013 sci-fi romance Her portrayed its idealised AI heroine (voiced by Scarlett Johansson) as rising up in another sense, ultimately shifting to a higher dimension where her human lover (Joaquin Phoenix) cannot follow.
Among the most complexly ambivalent of all depictions of AI is the 2001 film A.I. Artificial Intelligence – a project originally developed by Kubrick, and filmed by Steven Spielberg after Kubrick’s death. The protagonist David (Haley Joel Osment) is a robot “child” adopted into a human family, then discarded when circumstances change. With the innocence of Astro Boy and the relentlessness of a Terminator, he sets out to regain his lost paradise, which means tracking down the Blue Fairy out of Pinocchio and getting her to change him into a “real boy”.
An outlier in Spielberg’s career, A.I. Artificial Intelligence weaves familiar themes into a design too disconcerting for mainstream pop culture to imitate closely. Still, Ishiguro, an avowed cinephile, almost certainly had it in mind when he wrote Klara and the Sun – and perhaps the Austrian filmmaker Sandra Wollner did too in her genuinely shocking 2020 The Trouble With Being Born, which depicts an AI outwardly resembling a young girl (Lena Watson) living in a house in the woods with a paedophile (Dominik Warta).
To all appearances, the AI in Wollner’s film isn’t sentient or conscious – but given the expectations acquired over a lifetime’s viewing, it’s hard to feel absolutely certain. The same uncertainty arises in Alex Garland’s prescient 2014 film Ex Machina, which evokes the familiar sensation of wondering if we’re dealing with more than a machine.
Once again, the premise has the simplicity of a fairytale: an everyman programmer (Domhnall Gleeson) is invited to the compound of a brilliant, reclusive tech CEO (a bearded Oscar Isaac, looking a little like Kubrick). Here he meets Ava (Alicia Vikander), a robot who can pass for a real woman, and is invited to test if she’s truly conscious, perhaps as a cover for whatever trickier game the CEO is playing.
This is yet another Pygmalion story – and once we start to suspect that Ava may have hidden motives of her own, we’re also not far from the tradition of Metropolis, with its robot femme fatale. In this light, it’s possible to read the film less as a statement about technology than as an allegory of the gap between men and women (like Garland’s 2022 horror film, Men).
Still, by the end of Ex Machina there’s no doubt a new force has been unleashed upon the world. We can’t know what it will amount to, any more than we can be certain in 2026 of how AI will evolve. But if we assume that any creation reflects the nature of its creators, it follows that we’ll be getting no more (or less) than we deserve.
Klara and the Sun is in cinemas from October 22.
More from Spectrum
- Jason Donovan on Kylie, his dad, and wrestling ‘a monster’
- What a late-night tweet taught Ai Weiwei
- Kyle MacLachlan’s journey into the darkness
- Troye Sivan takes stock