How the Machines Will Kill Us All, According to The Movies

By  · Published on April 16th, 2015

Marvel Comics

This article is part of Humanity and the Machine, our exploration of the cinematic interactions between humans and self-aware machines.

If you’re reading this, there’s a good chance that robots haven’t yet eliminated the last dregs of human society. Congratulations! (unless you’re one of a handful of future war survivors, reading this on a relic PC in some bombed-out library – and don’t you have more pressing issues to attend to?).

But the fear of a coming robopocalypse is real. Very real. At least, that’s what the movies tell us, because it seems like a cinematic A.I. can’t become sentient without at least considering the possibility of turning on its human creators. For every A.I. with the genteel soul of Robin Williams, there are at least ten with glowing red eyes and bits of crushed human still wedged between their toes.

Kinda sucks to be on the human end of that. Which is why we’ve put together this guide, illustrating the potential futures that Hollywood tells us are coming when computers finally get sick of doing our dirty work. I know I wouldn’t want to spend all day shuttling emojis back and forth between people’s phones.

20th Century Fox

A.I. Apocalypse #1: KILL ALL HUMANS (And Mutants)

This is the gold standard for machine uprisings. A.I. is cold, unemotional, unmoved by all the polite gestures people do for each other every day. If one got the chance at Earth’s #1 slot, it’d take it in the least compassionate way possible: blowing us all away. Some robots even have sex fantasies about it.

“Please don’t kill me” seems like society’s default opinion on sentient A.I. The most famous cases of cinematic A.I. gone rogue- HAL 9000, Skynet (and by association, all non-reprogrammed-for-heroism Terminator models)- are famous for the cold machine deadness they exude when ending human life. Ditto for any recent A.I.-gone-rogue blockbusters, because as X-Men: Days of Future Past and Avengers: Age of Ultron warn us, one typo in the programming code and suddenly a mechanized army is slaughtering all human life.

That’s the one underlying value in all these “kill all humans” movies: computers are faulty and not to be trusted. Chalk it up to humanity’s fear of the unknown (or maybe humanity’s fear of killer robots). HAL 9000’s got a few bugs, and he hears the Discovery One crew consider shutting him down because of them. Thus: “kill all humans.” Days of Future Past’s mutant-exterminating Sentinels were programmed a bit too efficiently and eventually rationalize that humans will on rare occasions give birth to mutants. Thus: “kill all humans.” Skynet’s is hilariously simple. Humans create A.I., realize they made it a bit too powerful and try to shut Skynet off. Skynet says “no.” You can probably figure out the finer details from there.

Avoiding a “kill all humans” future is easy. Don’t build a sentient computer. Just don’t do it. Too much risk involved. And if it’s already been built (and is well on its way to depopulating the Earth), definitely don’t try appealing to its sentimental side.

Warner Bros.

A.I. Apocalypse #2: ENSLAVE ALL HUMANS

Kind of a sister apocalypse to the “kill all humans” thing, plenty of power-mad A.I.s have decided that there are benefits to keeping humanity alive and under machine control. Obviously, The Matrix is the gelatin-slick poster child for human enslavement, but you’ve got other A.I.s out there pulling the same scheme (less iconic, of course, given their lack of goo). I, Robot’s VIKI wanted to rule over mankind for its own good. Tron’s MCP felt like extending its control from a single computer mainframe to the Pentagon.

“Enslave all humans” is a significant upgrade from “kill all humans.” Killing all humans is easy. Relatively, anyway (it’s not like I’ve ever tried). Drop a few nukes, release a few robot armies and the damage is done. Enslaving humanity requires way more legwork, like building an infrastructure and devising some way of policing all those captive humans.

Complex plans = complex motives, which is why the A.I. that go for enslavement are usually more complicated characters. The machines of The Matrix, for example, with their tendency to launch into long, dense monologues about hacker philosophy and fate. They’re developed enough to use the word “concordantly,” a word our spellchecker doesn’t even recognize, which should be proof enough (also, actual proof: the two “Second Renaissance” segments of The Animatrix, which were all about sympathetic A.I.).

So if at some point in the future you find yourself the unwitting servant of a mechanical overlord, you might consider talking your way into freedom. That’s more or less what Neo did at the end of the last Matrix movie.

Columbia Pictures


A.I. is evil, computers spell certain doom, run screaming in terror from the Apple display at Best Buy. We get it, movies. But once in a great while, a computer will gain sentience and defy all odds by being kind of a cool guy.

From two recent films, we have two potential end-of-humanity situations that might actually be to our benefit. One would be Transcendence, where an A.I. Johnny Depp rigged up a cyber-utopia out in the desert, free from hunger, disease, pollution, sadness, etc. In essence, world peace. Depp-bot’s fatal flaw? Marketing. His perfect world involved pools and puddles of sludgy, metallic brown nanobots and human worker drones that, while the picture of perfect health, ran with the same “stiff arms, no blinking” ferocity as Robert Patrick’s T-1000. Can you really blame humanity for saying “no thanks” to a utopia with all the outward appearances of T2: Judgement Day?”

Extrapolate the ending to Chappie, meanwhile, (Spoilers for Chappie incoming) and you’d end up with a society that was also disease-and-hunger free, because everybody’s consciousness is uploaded into a robot chassis. Which could be a major win for humanity. But would probably require another crack marketing team, given how eerie Robo-Yolandi looks in Chappie’s final sequence.

When faced with the potential for a positive robopocalypse, proceed with caution. Sure, the A.I. offering you peace and salvation could really mean what it says. But for all you know, there’s a glitch in its CPU and salvation involves a lot of spinning buzz saws.

Warner Bros.

A.I. Apocalypse #4: EMULATE ALL HUMANS

All these situations involve one key ingredient: an apocalyptic event specifically triggered by artificial intelligence. What happens if humanity renders itself spectacularly extinct by its own hand, with the A.I. innocently watching from the sidelines?

According to Hollywood, the computer survivors would probably just keep on keepin’ on. Our best example is A.I.: Artificial Intelligence, which spends most of its time in the late 21st century as humans and Mecha (A.I.’s cutting-edge humanoid robots) live on the shaky cusp of human decline and robot prominence. At least until the last few minutes where, one ice age later, we end up in the late 41st century. Humans are gone and Mechas are the new dominant life form (also, Mechas have evolved from synthetic humans into sparkly grey aliens for some reason). They’ve got science, civilization, the works. At least, from what we can see. Humans are gone and A.I. just took over where we left off. As far as apocalypses go, that’s not so bad.

For A.I. that’s not quite so developed, another fate awaits. This would be a WALL-E situation (or a Silent Running situation if this was 40 years ago). Humanity’s gone, and whatever little shreds of A.I. remain just keep doing whatever they were supposed to. In WALL-E, waste management. In Silent Running, outer space gardening. Either way, the A.I. will keep it up, maybe until it finds love but probably until it drops dead after several thousand years of crippling loneliness.

Disaster-planning-wise, I’d suggest arranging all your various A.I. helpers in pairs so no one’s alone after we’re all dead. Well, until one of the pair dies. There’s really no happy ending here.


A.I. Apocalypse #5: ROMANCE ALL HUMANS

Ok, one caveat- none of the movies that fall into the “A.I. learns to love” category actually involve an A.I. going rogue and triggering the apocalypse. There may not be a single film out there where robots engage in gentle romance and also claim a decisive victory against fleshy human scum (Hollywood should really get on that). However, the future of humanity will certainly be eclipsed by robosexuality at some point (it’s gone from being a Futurama gag to something people are actually discussing with a straight face), and that’s almost an apocalypse. Might as well throw it in.

What have the movies taught us about artificially intelligent love? It can be between two robots- see WALL-E for the best case scenario, and Heartbeeps for the absolute worst. Love can also flourish between human and A.I., but cinema has taught us these relationships aren’t exactly built to last. Think of Theodore Twombly and Samantha in Her, and how that whole “singularity” thing really came between them at the end. Similar deal for Lisa in Weird Science, although without the singularity. Ex Machina, too, although I’ll leave that as spoiler-free as possible.

But don’t let that stop you from pursuing your robosexual passions. If the future of romance looks a little bleak, search for substance in cinema’s one eternal, boundless love between man and machine.