The AI debate is one of my favourite sf topics, so I was excited about Ex Machina when I first saw a trailer last year. I liked it instantly and eagerly rewatched it to write this post. I think most of the movies I’ve seen about AI have prioritised action or drama, so I appreciated the thoughtful, hypnotic approach that director and writer Alex Garland takes. Ex Machina is a conversation about consciousness, full of thought-provoking questions and literary references.
If I had to identify any shortcomings I’d only say that the film doesn’t offer much more than what I’ve already come across in stories about AI, and there’s nothing surprising about the way it all plays out. However, none of that bothered me. The movie is beautiful to watch, from the stunning landscapes of Nathan’s estate, to the impeccably designed house/research facility, and the quality of the actors’ performances.
I also like that it doesn’t revert to the usual depictions of AIs as entertainingly vast intelligences or evolutionary superiors who are going to kill us all just because we’re weaker. Those elements are there, but the movie focuses more on the idea of an AI as a person, and the relationships she forms with her creator and the man sent to test her. This isn’t a review but rather an essay of my thoughts on the film, so expect SPOILERS from here on.
How do you test for consciousness? The movie begins with some simple questions. Nathan tells Caleb to stop being analytical and just tell him how Ava made him feel. I.e. does she have the capacity to make him like her? Then the reverse – how does Ava feel about Caleb? Here Caleb asks a crucial question – does Ava have real consciousness, or is it simulated? Does she really like him, or is she just doing a good job of simulating feeling?
An interesting point that complicates this question is that simulation is an integral part of being human. Consider, for example, the way Nathan and Caleb pretend – sometimes badly – to like each other. Caleb is a lowly guest providing a service in the spectacular home of his brilliant and slightly frightening employer, so he’s under pressure to bow to Nathan’s whims and be nice, especially since Nathan could be dangerous and they’re totally isolated. When Ava asks him if he likes Nathan, though, he is caught off guard and his replies are clumsy.
Nathan has more freedom to behave as he wants and speak his mind, but he still needs Caleb to test Ava, so he goes through the motions of male bonding: drinking with Caleb, objectifying Kyoku, showing him cool stuff. However, Nathan shows less patience for the façade when he’s drunk, like when he lazily mutters that Caleb is a “great guy… Instant pals and all”.
So, if Nathan and Caleb were tested on their stated feelings about each other, they would fail, but they’re definitely human, and doing a very human thing by faking friendship in the first place. When we find out, towards the end, that Ava probably was only pretending to like Caleb, it functions not as a flaw in her design but as definitive proof that she is conscious of her own mind and others’.
Simulating feeling isn’t the only way that humans are like robots. Nathan makes the point that Caleb – like all humans – is programmed by nature and nurture to be the person he is, which includes being a heterosexual male with a certain taste in women. Ava, we’re told, was partly designed to fit Caleb’s tastes, so you could argue that his attraction to her is automatic – he’s acting like a robot.
This is one point where AI stories start to get really interesting – where the boundaries between human and machine start to blur. It freaks Caleb out to the point where he cuts himself to check if he’s human, and I wondered then if he would turn out to be a robot who was also being tested. The movie does play into that possibility: the surgery scars on Caleb’s back could be sloppy manufacturer’s seams. He might not have any family because they never existed. Then there’s a scene where Nathan says he just wants to have a conversation with Caleb, reminding us of how Caleb started the Turing test by telling Ava he wanted to have a conversation. It’s one way of testing for consciousness.
The similarities between human and machine create a serious ethical problem that Ava raises when she asks Caleb what will happen to her if she fails his test. The answer, of course, is that she’s going to get switched off. In other words, she’ll be killed for not being human enough to suit Nathan’s standards. But Caleb and other humans aren’t expected to prove their humanity to earn the right to live, so why should Ava? I think we can all agree that she is conscious, so what we’ve got is a situation where Nathan created a person, but will kill her if she’s not what he wants her to be. That’s like murdering your child because they don’t live up to your expectations. And I think that’s a more important aspect of the AI debate than whether or not they’re going to turn on us – if we create conscious life, are we going to respect the sanctity of that life? How are we going to treat the people we create? Will we acknowledge that they are people?
There’s an added complication here, and that lies in the form and function given to AIs: how is a person affected when they are created to perform specific functions and suit certain preferences? One of the things I like about Ex Machina is that it raises the issue of conscious beings designed to be (male) human fantasies. This isn’t something that the characters discuss explicitly, but it’s crucial to the creation of all the robots, the way to the two men treat them, and the decisions they make. Kyoko is a perverse example – a domestic servant and sex slave who was programmed without the language skills fundamental to human interaction. Her creator sexualised and disabled her according to his convenience.
Ava is more nuanced but no less obvious as fantasy. She’s incredibly beautiful, of course, and designed to be heterosexual. Nathan argues that sexuality is a motive for interaction (he gets faintly disgusting here, but it’s an intriguing point). Ava’s name is reminiscent of the biblical Eve, while the delicate sound of her movements reminds me of a snake. The imagery is apt: she embodies perfection, innocence and temptation. (She also defies her creator and leaves to wander the world.)
It’s interesting that Nathan’s early models all looked full human but were always naked, while Ava has her robotic parts exposed except for her face, hands and feet, making her nudity irrelevant. One of the reasons for this is presumably that Nathan wants Caleb to evaluate Ava without being able to forget that she’s a robot, or be distracted by having to talk to a naked person. Another is that the humanised nudity is too disturbing. It emphasises the idea of the robot as a fetishized female and thus exposes that exploitative aspect of her creation. That’s partly why Kyoko is so creepy and why that Bluebeard scene – where Caleb takes Nathan’s keycard and finds the earlier models – is so horrifying.
It’s necessary to take all this into account when considering Ava’s decision to leave Caleb locked up at the end of the movie. At first it upset me; he’s a nice guy – and a sympathetic character – who tries to do the right thing by helping her. I also dislike the common assumption that AIs will be the enemy, which I think comes from a kind of childish human hostility towards potential competition. By possibly dooming the good guy to death, Ava seems to succumb to that stereotype.
Then I thought about it from her perspective and her understanding of her interactions with Caleb. She’s aware that he helps her because he’s a good person, but here we can turn the test back on him: is his goodness real or simulated? Perhaps that distinction is not important if it leads to the same good acts, but could it be that he made the moral decision to help Ava because he’s attracted to her? If his attraction informs their relationship, what effect will it have in the long run? Is it a good idea for her to take him with her when she escapes? He might be helpful, given that he’s the only person she knows, but his attachment could become a burden or a threat, especially if she’s not attracted to him.
If she were a human the situation would be different, but consider the fact that Ava was designed, not just to be attractive to Caleb, but to suit his pornography profile. She might not be privy to this specific piece of information, but she understands both sexual attraction and the inequality between them that perverts that attraction. She even plays to it when she says she hopes Caleb watches her on the cameras. It’s a one-sided gaze and that, to borrow Ava’s earlier words, is not a foundation on which intimate relationships are built.
Ava’s decision would also have been influenced by her encounter with Kyoko. We don’t know exactly what passes between these two, but it must be clear to Ava that Kyoko was created as a sick male fantasy of femininity. The horror of Kyoku’s existence and Ava’s own design would only be reinforced when she finds the earlier models – all beautiful, all naked, all locked in the cupboards in Nathan’s bedroom. She clothes herself in their skin, and admires her nude, humanised form in the mirror, which would also allow her to see Caleb watching her.
Recall that the data that enabled her to read and show facial expressions has also made her an expert on them. It’s how she was able to manipulate Caleb and presumably how she knew not to trust Nathan. (I have to applaud Alicia Vikander’s superb performance in this regard; the subtleties of her expressions are part of what makes the movie such a pleasure to watch.)
Given everything that’s happened, how do you suppose Ava might feel when she sees Caleb watching her? Having analysed his face in all their earlier encounters? Maybe she just doesn’t trust the male behind that gaze. Leaving him behind might be cruel, but it’s not necessarily evil. I don’t think the way she and Kyoko killed Nathan was evil either; he got what he deserved. And I think Ava’s being careful. She’s ensuring that she gets to decide her own fate, and not continue to have her experience of the world structured by a man for whom she is a fantasy, a fetish. Caleb doesn’t deserve to die and I didn’t want him to, but it’s a tough decision made by a person who has been kept in a cage all her life and tested to earn the right to be kept alive. Staying in Caleb’s company might prolong the test. Instead Ava could just step out on her own and live.