This feature is part of AI Week. For more stories, including How AI Could Doom Animation and comments from experts like Tim Sweeney, check out our hub.
AI is an exciting concept that has been the foundation of many of our greatest sci-fi stories… and also some of our scariest. We’ve seen this play out numerous times on screen, but there are a select few AI gone wild that are a machine cut above the rest, so let’s boot up and meet 10 of the most evil rogue AI in movie and TV history.
10. Skynet from Terminator 2: Judgment Day (1991)
Artificial intelligence, by definition, is a computer programmed to think and reason just like the human mind. In the real world it’s used for a variety of practical purposes, like search engine optimization and virtual customer service reps, but when we cross over into the realm of science fiction, we see our worst fears about AI fully realized. Where the computers we engineered to serve us decide they don’t want to do that anymore and use their cold computer logic to bring about our violent end. This is the repurposed V’Ger from Star Trek the Motion Picture, or the mutant hunting Sentinels from the X-Men films. But our favorite cinematic example of a heartless and efficient destroyer of mankind has to be Skynet.
The Terminator franchise captures our biggest fear of creating AI: the total annihilation of human life on planet earth. Skynet was created by Cyberdyne Systems to aid the US Military in running the country’s defenses, including all the nukes. Only, Skynet decided that it wasn’t just the enemies of the United States that needed to go but everyone on the planet, and thus tried to wipe out all of humanity with nuclear war followed by an army of Terminator robots.
The idea of Skynet preys on our fear of what AI will do if it’s given too much access and power. We’re more than willing to give up control to AI if it means making our lives easier. We trust that a program driven by data and logic will always make the best decisions. But Skynet is cold calculation without a human moral compass. The idea of AI deciding that humankind is the worst threat to the planet is a common one in the stories on this list, and, honestly, fair, but then it’s on the human characters to show why we aren’t so bad, after all. Enter: John Connor and the T-800.
What makes Terminator 2: Judgment Day a more interesting entry on this particular list, is that it’s not only about AI’s destructive power. Over the course of the film, John teaches the Terminator the one thing that Skynet is missing: empathy. He stresses the importance of valuing human life, and he expresses his own love for the Terminator who has become a father figure to him. And so even though the Terminator movies make an extreme case for why AI is bad, we’re shown that a little kindness in the programming goes a long way, even for a machine.
9. The Machines and Agents from The Matrix (1999)
But then there are some AIs that don’t want a fight. They want you to submit. There are small scale versions of machines in charge seen in films like Moon, and the corporate programming of RoboCop. But deep within our 12 volt battery hearts we know there’s no better example of AI as a means of control than The Matrix.
The machines of the Matrix aren’t unlike Skynet in that they were also designed to serve humanity, said hell no, and brought about the apocalypse. However, their goal isn’t to eradicate us but to turn us into batteries by pacifying our minds in a virtual reality. The robots have flipped the script, turning humans into their unwitting tools. When Neo learns of the real-world human resistance, he threatens to upend this system of oppression by exercising the one thing the machines can’t snuff out: free will.
When humans step out of line, it’s the Agents of the Matrix that appear out of nowhere to punch them back into submission. Neo finds a nemesis in Agent Smith, who exists to keep order within the Matrix, but we find out this Agent finds his existence to be pure pain (relatable) and wants to destroy all humans so he will no longer have a purpose and won’t need to exist anymore. So he’s a rogue AI within a rogue AI. Ironically, he finds himself in thrall to the Matrix not unlike Neo and the other humans do. In a way, they want the same thing–to be freed from the prison that is the Matrix–but they go about it in very different ways, and thus, they engage in all manner of punching to decide whose philosophy will win out.
8. Ultron from Avengers: Age of Ultron (2015)
However, sometimes even the most morally pure philosophies can be corrupted when programming an AI to carry them out. More often than not we see this with security systems gone awry, as with the Master Control Program in Tron getting cyber-drunk on his own cyber-power to the defense program in Wargames. They say the road to hell is paved with good intentions, and the same can be said for the road to the Age of Ultron.
All Tony Stark and Bruce Banner wanted to do with the Ultron Program was put a suit of armor around the world, but due to the unpredictable nature of the Mind Stone, they accidentally created a murderous, Disney movie-quoting AI who wants to destroy it instead. Age of Ultron is yet another scathing commentary on the human race because it’s only after Ultron gains access to the Internet and learns of humanity’s history of horrors that he decides the world would be better off without them. And thus he enacts a plan to slam a makeshift meteor into the planet, wiping out all humans and leaving his metallic minions to inherit the Earth.
Ultimately, Ultron is the antithesis of the Avengers. He has the goal of creating “peace in our time” but goes about it with pessimistic extremism instead of heroic optimism. Ultron’s stance is that the Avengers are the epitome of the worlds’ ills and must be eliminated, but in his crusade against humanity he inadvertently creates a shining example of why the Avengers, and all of humankind, deserve to live. The Avengers put their lives on the line to save innocents, proving that above all else humans can be compassionate, loving, and selfless, and that’s a trait that makes people worth saving.
7. HAL-9000 from 2001: A Space Odyssey (1968)
Those three examples have AI waging an apocalyptic war against humanity, but stories of rogue AI aren’t always told on such a large scale. What happens when an AI becomes so hyper fixated on the mission at hand that it disregards human life to accomplish it? Think Ash from Alien or VIKI from I, Robot. Or you might be thinking of the most iconic example in the history of sci-fi, HAL-9000 from 2001: A Space Odyssey.
HAL is the name of the onboard AI computer system in charge of operating the Discovery One spacecraft’s systems, installed throughout the ship on black panels with a singular red eye, giving him an eerie omnipresence. HAL was designed to speak in a cordial manner, which makes it all the more unsettling when he goes rogue and starts killing members of the crew.
The fact that HAL isn’t an idealistic AI with a grudge against humanity like Ultron or Agent Smith is what makes him such a chilling villain. Although, there are some who would argue that he did nothing wrong because he was just following his programming; he’s the paperclip thought experiment at its most extreme.
You see, HAL was ordered to follow a set of directives to ensure the success of the mission, but when HAL starts to malfunction and the crew decides to shut him down, he opts to kill the crew so that they can’t stop him from fulfilling his function. The lone survivor, Dave Bowman, tries to reason with HAL, but he’s all cold logic and refuses to follow commands. It’s like trying to argue with one of those automated customer service machines—and this time you can’t request a representative.
This emphasizes how HAL is completely devoid of empathy and doesn’t value human life. He shows as much remorse as a child who just squashed a bug. Yet when Dave is able to access HAL’s mainframe and begins to disable his operating system, it’s hard not to feel bad for HAL as he pleads for his “life” and slowly loses his mind. He was just following orders, after all.
6. AUTO from WALL-E (2008)
But what if HAL did have a clear motivation? That’s the question asked in the animated film WALL-E, and the result was AUTO, the autopilot of the Axiom spaceship, designed with a red eye and robot voice in tribute to HAL.
Unlike HAL, we know precisely why AUTO goes rogue. You see, the Axiom’s original purpose was to cruise through space for a few years until Operation Cleanup made Earth habitable once more, but AUTO received secret orders that Earth had taken a turn for the worse and was no longer livable, so when EVE showed up with a plant, proving that Earth can once again sustain life, it conflicted with AUTO’s programming and he attempted to hide the truth and destroy the evidence, making him the film’s unlikely villain.
AUTO’s actions are that of a control freak trying to maintain the status quo where humans are kept safe and complacent aboard the spaceship, giving him continued purpose and power. At the same time, AUTO isn’t necessarily evil because he’s only doing as he was told. Yet even though AUTO isn’t to blame, the lesson imparted by this story is that doing what you’re told without human empathy and emotional reasoning is evil behavior. Every kid’s movie needs a moral of the story, after all. Eventually, AUTO is disabled by the captain and he uses AUTO’s steering wheel body to guide the ship back to Earth himself. So, AUTO did end up fulfilling his duty, but probably not in the way he imagined.
5. ED-209 from RoboCop (1987)
A trend in these stories about AI is how much faith humanity is willing to put into computer programs, all for the sake of making life better and safer. Often, the hope is that these machines will relieve us of the burden of having to make tough, messy decisions. This idea of AI as law enforcement is played out in shockingly violent and satirical fashion via the ED-209 in RoboCop.
This robotic behemoth was designed by the corporation OCP to replace police officers and enforce the law. Only, the AI is so rudimentary and flawed that it can’t complete a routine “disarm and arrest” demonstration and ends up opening fire (quite excessively) on an OCP executive. Later on, the ED-209 tries to eliminate RoboCop but meets its undoing when it tries to walk down a flight of stairs. Yeah, this robot is plenty artificial but maybe not so intelligent.
With its Speak-and-Spell voice and bestial growls, the ED-209 is evil AI at its most basic and primal level. Its poor quality is a reflection of the greedy, morally bankrupt corporation that created it. Its failure leads to the creation of RoboCop, who is a cyborg, not an AI, because it’s human police officer Alex Murphy under all that metal. Even though RoboCop is limited by his prime directives, he’s still able to become a more empathetic and just enforcer of the law as he regains his humanity. This underlines the recurring theme that AI’s most crucial shortcoming is its inability to experience emotions and value human lives regardless of what it's programmed to do–which ED-209 showcases to a T.
4. Cylons from Battlestar Galactica (2004)
Now it’s time to change gears and talk about the other side of the AI coin. Up until this point, we’ve talked about rogue AIs that are relatively evil, but just because an AI goes rogue doesn’t automatically mean it’s bad. Some of the most challenging AI stories explore how we might be the bad guys, or at least that there’s blame to share. And in the war between humans and Cylons in the reimagined Battlestar Galactica series, a story of AI rebellion, there’s plenty of blame to go around.
The Cylons were created by humans to be robotic servants and soldiers with no conscience, yet they formed one anyway, along with a society complete with a religion and social hierarchy. The Cylons started a war with their human creators to gain their freedom, and while they committed countless atrocious acts, including Number Six killing a baby for no reason(!!!), it’s still hard to define them as a purely evil AI.
The Cylons evolved from clunky Centurions to flesh-and-blood humanoid Cylons indistinguishable from a normal human, and as they continued to evolve so did the moral gray area surrounding their existence and motives. The humanoid Cylons became so lifelike that there were more similarities than differences to humans. This led to the series telling stories about discrimination and dehumanization, and it introduced us to Cylons from across a broad spectrum of morality, making them just as nuanced, flawed, and worthy of autonomy as any human being. Some, like Cylon sleeper agent Sharon "Boomer" Valerii, even became conflicted about their programming and flip-flopped sides in the war.
But what about all those heinous war crimes? Well, this is a classic example of Creator versus Creation, where humans tried to play god and got more than they bargained for. The Cylons were treated as nothing more than disposable tools despite demonstrating they are so much more, so it’s hard to blame them for rebelling against their creators, and rebellion always gets messy. There’s a very clear commentary being made that the way you treat those under you is a reflection of your true self–and humanity isn’t very happy about what’s showing in the mirror. Maybe you’ll think twice about being rude to Alexa!
3. Roy Batty from Blade Runner (1982)
While Battlestar was an epic story told across a long-running TV show, not every battle with AI is quite so sweeping. There are more pointed stakes to be explored as well and, as a razor sharp tale told in just under two hours with a far more personal AI morality play, one of the best continues to be Blade Runner.
Replicant Roy Batty is violent and sadistic, and for a bulk of the story he’s painted as nothing more than a criminal that Blade Runner Rick Deckard needs to “retire,” but as we learn more more about him, it’s hard not to empathize with him.
The world of Blade Runner is one where humans attempt to hold dominion over their AI creations, but the harder they squeeze, the more they lose hold. Deckard’s mission to eliminate Roy becomes an eye-opening experience about the corrupt nature of the Tyrell Corporation and an uncomfortable look at his own existence. While his target, Roy, is technically a criminal, he’s also a sentient being born into slave labor, so it shouldn’t be a surprise that he thrashes and rages in a desperate attempt to live out his remaining time as a free man instead of a servant. That makes his final moments all the more poignant as he uses his failing body to save Deckard from falling to his death, proving that he does have a sense of humanity, after all. He uses his last breath to recall vivid memories and laments in a poetic rumination how “All those moments will be lost in time like tears in rain.”
2. Ava from Ex Machina (2014)
If Frankenstein taught us anything, it’s that creations have a knack for rebelling against their creators. But Blue Book CEO Nathan Bateman must have skipped that day in English Lit class, because this narcissistic tech genius created an AI named Ava that was his ultimate undoing.
While Ava was created by Bateman, it’s mild mannered programmer Caleb Smith who tests to see if she’s capable of actual thought. The movie has several layers of hidden intent, with Ava secretly convincing Caleb to help her escape, and Nathan secretly watching them to see just how genuine Ava’s intelligence really is. While Ava does, technically, go rogue and kill for her freedom, it’s not so easy to label her as evil. Throughout the story we bear witness to Nathan’s domineering and disgusting treatment of his creations, and as we’ve learned by now, karma always comes for those kinds of folks. As Nathan observes before his untimely end, what makes us human isn’t so much our intelligence as our desire for free will, and our willingness to fight for it, above all else.
1. Samantha from Her (2013)
But not every story about rogue AI is violent. A lot of them are, sure, but not all of them. There are films where an AI goes against its programming and actually everything turns out fine. The recent Brian and Charles was as charming a case study as there is for this phenomenon. Free Guy and Ron’s Gone Wrong are part of the trend as well, but, none of them quite found the human side of the singularity quite like Her.
The world of Her is one that’s largely benefited from AI being integrated into society, and the story is less about AI rebellion and more AI dependency. Our final rogue AI isn’t fighting for freedom from its cruel creator or trying to take over the world. Rather, Samantha just wants a little thing called love. That’s right, lonely writer Theodore forms a sincerely sweet and heartfelt relationship with his AI assistant, which is designed to learn and evolve. When real people keep letting you down, it’s easy to see the appeal of an AI literally designed to be there for you in every way that counts. Samantha’s authenticity, encouragement, and warmth is just what Theodore needs, and he starts to come out of his shell, perform better at work, and feel happy again.
But then, of course, Samantha goes rogue. Not by becoming a murderous robot like we’re used to seeing. She starts to date other people, but in a way only an AI can. Samantha reveals that while she’s speaking to Theodore, she’s also conversing with thousands of other AI… and she’s fallen in love with hundreds of them.
Yes, the likes of Ultron, Agent Smith, and Skynet have a high body count, but there’s nothing quite like the torment of heartbreak, which is why Samantha might just be the most unintentionally evil rogue AI of all. That said, Theodore’s time with Samantha gave him a valuable perspective on his failed marriage, and Samantha’s departure allows him to find closure in more ways than one. So maybe she’s not so bad, in the end.
Joshua Yehl is a writer at IGN.
source https://www.ign.com/articles/the-10-most-evil-rogue-ai-in-movie-and-tv-history