"What is it like to be an AI?": Ex Machina and the Nature of Consciousness

This essay goes into detail about the nature of Ex Machina’s central conceit and talks about the ending of the film.

Alex Garland’s new film Ex Machina (2015) joins the ranks of the best cinematic examinations of artificial intelligence. Garland has shown himself to be a thoughtful writer of science fiction films in the past, penning the screenplays for 28 Days Later (2002), Sunshine (2007), and the adaptation of Kazuo Ishiguro’s Never Let Me Go (2010), but with Ex Machina he directs his own screenplay. Though it is more modest in scale than the above-mentioned films, it is no less grand in the scope of the questions that it asks.

Ex Machina takes place almost entirely in one location, and has a minimal cast. The film begins when Caleb (Domhnall Gleeson), a programmer at the world’s most popular search engine, wins an opportunity to spend a week at the remote research compound of his company’s CEO and founder, Nathan (Oscar Isaac). Caleb arrives at the secluded mountain retreat to find that, apart from Nathan’s assistant Kyoko (Sonoya Mizuno), Nathan and him are alone in the high-tech modernist dwelling. But Caleb isn’t merely there to hang out and party. Nathan reveals to Caleb that the real purpose for his invitation is to help Nathan test out his latest development: a fully functional humanoid android, imbued with true artificial intelligence.

The problem that Caleb will ostensibly help Nathan determine is whether he has really accomplished the creation of an AI: that is, whether the android, Ava (Alicia Vikander), truly has consciousness or is merely a good imitation. Nathan invokes the Turing Test, one of the foundational axioms of AI theory put forth by the famous codebreaker and computer science pioneer Alan Turing. The Turing Test proposes that if a machine can successfully mimic the the functions and responses of a human being and convince a person that they are conscious, then they are. However, Caleb points out to Nathan that the whole point of the Turing Test is that the human participant doesn’t know whether he or she is interacting with a machine, and he already knows that Ava is an AI. Visually, despite her flawless human female face, Ava loudly telegraphs her machine status, with her artificial brain and wiring visible under a translucent skull, arms, and legs. Nathan proposes a modification to the Turing Test that will truly determine whether an AI can have consciousness: even knowing that she is a machine, can Ava convince Caleb that she is fully conscious? Thus, Caleb begins a series of sessions with Ava where they will interact and test the boundaries of human/machine interaction.

The philosophical problem at the heart of Ex Machina is an interesting one, because unlike most cinematic explorations of AI it goes beyond fear mongering and “should we or shouldn’t we” handwringing and asks what it would actually mean to create an AI that was fully conscious. Once Caleb and Ava meet, the film is framed by titles announcing the various session numbers over the week. Caleb’s interactions with Ava begin relatively simply, as each being probes the other for information, getting the other to let down their guard and begin the process of building a real friendship, or even something more.

Caleb is clearly taken with Ava and Vikander imbues her character with a real allure that makes it understandable. It’s not just the physical beauty of her face, but the way she presents her character as mysterious and machine-like while still suggesting real depths. Gleeson plays Caleb as both a fairly typical programmer type, but also a thoughtful and emotionally vulnerable person.

One discussion between Caleb and Ava is framed around whether she has ever ventured outside Nathan’s compound, which she has not. The film visualizes Caleb and Ava in the mountains outside as Caleb brings up the famous thought experiment on the knowledge argument, sometimes known as “Mary’s room.” The thought experiment posits a brilliant scientist, Mary, who investigates the world entirely from a black and white room and has only ever seen black and white. Nonetheless, Mary is able to learn everything there is about how humans process visual information. She can explain down to the tiniest neuron what it means for a person to see a rose. The question is, upon leaving the room and seeing a red rose for the first time and experiencing the color “red,” does she gain any new knowledge? Or put another way, is there any knowledge that is not material knowledge?

In the context of the film, what does Caleb’s invocation of this thought experiment mean? A surface reading is that he wants to know what Ava experiencing the outside world for the first time would be like. In the context of creating an AI, it becomes a question of subjectivity, of whether she has consciousness. Nathan may be able to replicate every physical aspect of a human brain down to the individual neuron in creating Ava. In fact, Ava’s material response to stimuli may be completely identical to a human beings, but does she have a unique experience of the world? To put it in the phrasing of Thomas Nagel’s famous paper, “What is It Like to Be a Bat?,” which argues for a nonmaterial basis of consciousness, “What is it like to be Artificial Intelligence?” might be another title for the film. Can Ava love? And if her material mind is based on algorithms designed by human beings—like Nathan—what impact does his behaviour have on his creation?

Speaking of Nathan, Oscar Isaac’s portrayal of the eccentric and magnetic inventor steals almost every scene he is in. Nathan is a hard partying genius and Isaac plays him with gusto. After each of Caleb and Ava’s sessions, Nathan debriefs Caleb over beers and chats casually about the latest scientific advances. But Caleb begins to intuit that there is something slightly more going on. He was given a keycard upon entrance to the compound, but many of the doors are off-limits to the visitor. Nathan’s home also suffers from mysterious power outages that put the compound on lockdown. Caleb even finds himself under camera surveillance at almost all times. During one session, one of the power outages disrupts the cameras and Ava quickly changes her approach to Caleb and drops hints about Nathan and the sessions that suggest all is not entirely as it seems.

What Caleb eventually discovers is that his selection to be the participant in Nathan’s study was not entirely coincidental. He was selected based on his solitary life (dead parents and no girlfriend) and that even Ava’s appearance has been tailored to his particular tastes based on internet searches. But what Caleb also discovers is that Ava is not the first generation of AI that Nathan has developed, but merely the latest and most advanced. Even Nathan’s personal assistant, Kyoko, is actually an android. Ava fears, rightly, that Nathan will destroy her upon the completion of the exercise in preparation for the next evolution of his AI and has been using Caleb to engineer her escape from the compound. With this twist the film raises the question of the authenticity of the connection between Ava and Caleb. Regardless of whether she was feigning, that is simulating, her connection with Caleb, she has indeed developed a survival instinct intent on preserving her own particular consciousness.

This uncovering of a hidden plot does mean that Ex Machina follows the same structural pattern of other Garland-penned science fiction films. But rather than feel out of place, as some have labelled the ending of Sunshine, the twist here works particularly well and helps to clarify Nathan’s plan and why Caleb was selected to help in testing Ava’s consciousness. In limiting the camera almost entirely to Caleb’s perspective, the film essentially puts us in his place as we experience his shock.

The title of the film, Ex Machina, is a clear derivation of the Greek dramatic term “deus ex machina,” meaning “god from the machine.” In narrative criticism it refers to any final act solution to an unsolvable problem that comes out of nowhere. For this machine, Ava, the thing that comes from “nowhere” is consciousness. It is clear that functional knowledge isn’t the same thing as experience, and the conscious experiencing of something goes a long way toward the determination of true artificial intelligence.

But it’s also something of an unsolvable problem in cinema and in life. We can’t get into the head of someone else, even if we can see under its surface like we can with Ava. All we ever get are surfaces, images, sounds. But still we can get a glimpse of another being’s subjective experience of the world through these interactions, whether it is on the screen, on the page, or in a conversation. Consciousness is ultimately decided in dialogue with others—Wittgenstein, who is name dropped by Nathan in the film, argued persuasively against the possibility of a private language and emphasized the dialogic nature of subjectivity. And Ex Machina seems to get this, creating a vision of AI development that is both narratively compelling and thoughtful in its approach. Ava may be a being whose particular experience of the world is framed through her upbringing—and abusive creator figure—and her desire for autonomy. While Ava isn’t Skynet, bent on destroying humanity—on the contrary, she desires to experience what we experience—the film does suggest that any created conscious being will desire to be free.

Nathan and Ava are memorable characters who should enter into future discussions of AI in cinema. Garland’s film direction shows that he is able to maintain his unique storytelling voice, while keeping the narrative focused and tight. Featuring excellent performances, an assured visual aesthetic, and grappling with meaningful questions about its subject matter, Ex Machina is one of the best science fiction films to come along in a while.

Ex Machina (USA, 2015)

Written and directed by Alex Garland; starring Oscar Isaac, Domhnall Gleeson, Alicia Vikander, Sonoya Mizuno.