AI has been a popular topic for movies and fascinated audiences for over 60 years. Hollywood hasn’t always gotten it right—in fact, most of the time, they’ve been totally off. What is it about artificial intelligence that entices directors? What are they missing when they create movies about AI? Let’s take a look at a few popular movies about AI, why the topic is popular, why some got it wrong, and a few movies that actually got it right.
[Related Article: Is AI the Future of Digital Influencers?]
The Allure of Movies About AI
To most people, artificial intelligence creates a question of what is and what isn’t human. The idea of creating intelligence makes people who don’t understand the limitations and true scope of AI go crazy with the possibilities of the future—possibilities which aren’t necessarily plausible. The allure of artificial intelligence is using it as a symbol for the things humans fear or desire, and if you don’t get caught up in the truth of the technology, they make for exciting and thought-provoking storylines.
Film is one of the easiest mediums of shaping public opinion, and by ignoring the real possibilities that AI technology can offer us filmmakers are deciding what the public believes, even if it’s false (which, in turn, could even shape public policy that affects our work). They’re also missing out on an entire realm of stories that showcase what it means to be human, in actual humans.
The Main Misstep
The place most filmmakers make mistakes is in the abilities they give their AI. Artificial intelligence, or any form of machine learning, is generally programmed to perform one task or a series of related tasks or functions. But directors usually make their AI’s abilities far too complex, or the AI ends up somehow deciding that it won’t listen to its programmed task.
The first 10 minutes of I, Robot actually got really close to what our society could look like when we integrate robots into our everyday lives: task handlers and personal assistants. Unfortunately, it got carried away out of the realm of futuristic realism when a supercomputer created her own rule to robotics, when robots generally decide not to carry out their programmed task, and when a robot, Sonny, has evolved to have consciousness.
The 2013 movie Her did create a task-focused AI operating system, who names herself Samantha, suggesting a possible future to our current day Siri’s and Alexa’s. However, she too completely ignores the boundaries of what we’re currently capable of designing. We aren’t just yet at the stage to make AI that functions like this in the near future, though steps are being made for it to become a reality eventually. But it’s unlikely that it’ll happen in a near enough future that the rest of our society seems to be the same (like it is in Her—notice those ‘70s inspired clothing and overly-similar Los Angeles skyline?).
Ex Machina completely ignores this idea of simple-task-focused design with its main character/robot, Ava, and instead seems to insinuate her “task” is to become human—or as close to it as possible. While we’ve gotten pretty far in making robots that mimic humans, we haven’t gotten anywhere close to artificial general intelligence or creating robots that have real feelings (like anger, love, or sadness, as Ava is shown to feel in the film).
There are a handful of other examples that push the envelope too far from what we know to be possible. From Transcendence and Tron with the idea of “uploading” your consciousness; to 2001: A Space Odyssey where an AI turns on those it’s supposed to help; to Bicentennial Man in which a personal assistant AI develops emotions and tries to make himself human, eventually getting married. Hollywood loves to question the “what-ifs,” even if they just aren’t possible, as is the nature of creating good narratives of fiction. A movie about a machine learning algorithm to run automated marketing campaigns just wouldn’t be as thrilling.
Which AI Movies got close?
While it is more common for movies to get depictions of artificial intelligence wrong, there have been a few instances that have gotten it right (or almost right). And, they’ve managed to do this while still creating thought-provoking narratives, questions of humanity, and possibilities of the future.
One of the most beloved AI-centric movies of all time is the 2008 Disney movie, Wall-E. It follows the title character, who is a trash compactor robot, on his journey off of Earth and into a spaceship designed to protect humans from the pollution they’ve created on earth. While it does give the main characters, Wall-E and Eve, emotions and decision-making skills, the movie is a great representation of what the world could look like when robots are fully integrated into our society. Most of the robots follow specific, simple tasks, and that’s all they do throughout the movie. There are cleaners, robocops, security guards (like this Cambridge company), and a ton of others, who all do as they’re programmed and (try to) help humans, as they were designed.
Wall-E questions the damage that human consumption and laziness creates, and showcases robots as both the cause and solution to those problems, like they are in real life.
Another great example of an artificially intelligent computer doing the task its programmed to do is the simulator “Joshua” in WarGames. The movie is set in the cold war, and follows a teenager, David, who hacks into a computer to try and play a game “ThermoNuclear War,” which is actually a computer that is programmed to start and win the war if a threat is received. When David “plays” the computer, it sets off its sequence to start the war, as soon as it can figure out how to win. Panic ensues, and eventually the computer is told to play against itself. Through reinforcement learning, the computer realizes that there’s no way to win the war, except to not start playing. The computer never goes beyond what it has been programmed to do, and it’s even a design that we could create today.
In WarGames, the characters are faced with an antagonist, the computer, who isn’t inherently evil, but is simply doing what it’s been programmed to do. It’s a great depiction of the limits to AI—the computer doesn’t realize that it’ll be killing millions of people and destroying entire cities, it only knows that it’s supposed to win.
While it’s not a movie, I couldn’t write an article about AI in media without talking about Black Mirror, and, specifically, its depiction of AI in the episode “Be Right Back” from Season 2. At some point in the near future, I’ll be writing an article entirely focused on Black Mirror and the way it uses technology to create narratives, so I’ll keep this focused on the AI.
This episode is about Martha, after her boyfriend, Ash, dies in a car crash. She’s introduced to a technology which can read all of Ash’s social media and create a simulation of him. It starts out as chatbot mimicking his writing, then she feeds it audio and it’s able to call her in his voice, and eventually she upgrades to a humanoid robot that looks, talks, and acts like him—to an extent.
There are a few gaps between what we’re able to do now and what is depicted in the episode: while we can use DeepFakes and machine learning to edit photos or videos and replicate speech, it has to be programmed or scripted by a human; while we have humanoid robots, the episode depicts a fully-functioning body that can generate its own appearance.
But overall, the episode stays true to the limitations of GANs and machine learning. Martha grows frustrated with the reproduced Ash because it’s only able to act based on how it’s learned from Ash’s social media, or based on how Martha tells it to—it can mimic what it’s already learned, not create new ideas. It also isn’t able to go within 35 feet from its activation point unless it’s accompanied by Martha—it can’t just run off and be its own person (like in Ex Machina).
Like in WarGames, this Black Mirror episode shows the limitations to artificial intelligence. While intelligent in its own way, AI doesn’t, and can’t replace humans and human emotions. They’re limited to what they’re taught and trained for.
[Related Article: AI Ethics: Avoiding Our Big Questions]
Fortunately, this isn’t a bad thing. AI has had massive success in taking over the menial tasks that crush human spirit (like depicted in Wall-E), and has been able to greatly reduce human margin of error in applications like drought management or healthcare services.
The Future of Movies About AI
Humans will always question what makes us human, and will always question what lies in the future. Whether that means getting the technology wrong and blurring the line of “artificial,” or getting it right and showing the separate places where human nature and robotic precision flourish, we’ll continue to get exciting, entertaining, and engaging media that pushes us to examine ourselves.
And, as AI becomes more powerful, it’s likely that the ethics of AI will become a more common theme as well. The further we go from “what-ifs” to “what is,” the closer we will be to having to address genuine concerns about AI and ethics.