10 Fictional AI, Ranked by How Likely They Are to Kill Us All

10 Fictional AI, Ranked by How Likely They Are to Kill Us All

Mrs. Davis is a genre-bending new take on artificial intelligence: Here is a TV series in which an action hero nun takes on an AI with dream of world domination. (If this is the first you’re hearing of it, well, it’s on Peacock, so.)

The show kinda pits faith against technology, though it’s not nearly that pedantic. Simone (Betty Gilpin) is a nun who spends her spare time debunking charlatans, and the first episode sees the all-powerful AI (known as Mrs. Davis) tasking her with destroying the Holy Grail. But Simone blames Mrs. Davis for the death of her father, and doesn’t trust the computer in the least. Cue techno-thrills.

Mrs. Davis is but the latest in a long line of artificial intelligence narratives that go back centuries, way past stories of sentient robots to all manner of anthropomorphic animated objects, often brought to life by magic. Those variations on the theme were typically parables, but as we get closer to creating an artificial intelligence that could actually pass for human, we’re coming face to face with moral quandaries that no longer feel strictly hypothetical. Grammarly has been policing our writing for a few years already; now other AI tools are writing entire articles. Now that AI has been used to create a new song from Drake and The Weeknd, we’re getting closer to the day when those in power will start to seriously consider whether they need real human workers at all.

When that happens, we can’t say the movies didn’t warn us — though not all fictional AIs are malevolent. Here are 10 AI from movies and TV shows, ranked by how likely they are to kill us all.

Person of Interest (2011 — 2016)

Think Minority Report, but with artificial intelligence standing in for the three naked precogs in a tub. Michael Emerson plays Harold Finch, a reclusive billionaire who designs and codes a system (called only the “Machine”) that collates enough information from global sources to predict terrorist acts and identify potential perpetrators (the show prompting the question: what if the world’s richest person wasn’t completely useless?). While operating, on a surface level, as the kind of CBS police procedural your parents love, the show (from Westworld creator Jonathan Nolan) asks fairly smart questions about free will and privacy (while generally leaning on the safety > liberty side of the argument). It’s more interesting than it has any right to be, given its relative success on network TV.

Malevolence level: There’s no real suggestion of the Machine’s sentience, and everything remains largely up to the humans…but we’ve seen what happens when law enforcement tries to anticipate crimes and criminality. 5/10.

Colossus: The Forbin Project (1970)

Designed by Dr. Charles A. Forbin, Colossus is a computer system built to control the nuclear arsenals of both the United States and its allies. Buried deep within a mountain and drawing on its own nuclear power source, the computer is meant to be completely secure from any outside attack. Unfortunately, the its own draconian defence logic leads it to conclude that humans aren’t capable of making their own decisions. Even more unfortunately, the Soviet Union has built a similar system and the two computers decide they’re on the same page, proclaiming that humans may have either “the peace of plenty and content, or the peace of unburied death.” Positing that freedom is an illusion anyway, the linked computers proceed to take complete control of world affairs, with a global nuclear arsenal to keep people in line.

Malevolence Level: While Colossus is very much the villain of the piece, it’s never entirely clear it’s wrong about our inability to manage our own shit. 8/10.

Upgrade (2018)

Leigh Whannell, better known for his collaborations with James Wan on the Saw and Insidious movies, crafted an impressively smart action thriller in Upgrade, a movie that plays a bit like a cyberpunk John Wick while asking big questions about the wisdom of incorporating AI into our daily lives. In the near future, artificial intelligence is everywhere. The plot kicks off when an AI-controlled vehicle malfunctions and crashes while AI-run drones monitor the entire situation (not particularly far-flung scenarios from a 2023 perspective). Luckily for our grievously wounded protagonist, Grey, there’s some new tech soon to come to market: STEM, an implant that uses AI to take over the work of his severed spine…and then some. With his newly enhanced abilities, he seeks revenge on the people responsible for the accident that injured him and killed his wife. Without ever sidestepping the action, the movie explores some heavy territory, re: personal responsibility in our brave new world. Grey and STEM become so interconnected that it’s hard to know where one ends and the other begins. In a word where we more consistently outsource our decision-making to machines, who’s responsible when bad things happen?

Malevolence Level: STEM becomes increasingly malevolent as the film goes on, though everything it knows it learned from us. 7/10.

Black Mirror: “Be Right Back” (2013)

Several episodes of the (usually) dystopian sci-fi anthology Black Mirror touch on artificial intelligence, but “Be Right Back,” which kicked off the second season, asks particularly timely questions. In this one, Martha Powell (Hayley Atwell) is grieving the unexpected death of her partner, Ash (Downhill Gleason) when she learns of a service that will recreate a virtual version of a dead loved one using videos, photos, social media posts — essentially leveraging our increasingly large digital footprints to reconstruct us once we’re gone. At first, she’s able to chat with “Ash” on the phone, losing herself in the fantasy, but then the company promises a more “physical” option. This might have seemed a distant possibility as recently as 2013, but the ubiquity of tools like ChatGPT have made this kind of thing more than just a possibility: companies are already doing it, at least in rudimentary form. It’s now an inevitability, and raises all sorts of complicated questions, the most disturbing of which are related to our value as distinct and finite individuals in a world where we can be replicated endlessly. Would retreating into a secondhand fantasy version of a relationship be dehumanising, or perfectly fine, if it makes us happy? Would you settle for an approximation of a lost loved one, or want one of you to linger after your death? Better start thinking about our answer.

Malevolence Level: Though his existence raises very disturbing questions, Ash himself seems entirely benign. 3/10.

Transcendence (2014)

Transcendence isn’t a great movie on a plot level, but it does (largely) play fair with the real potential of artificial intelligence, at least as it’s likely to evolve in the near future. Dying of polonium poisoning, a scientist (Johnny Depp) proceeds to image his brain and consciousness for upload into a computer, forming the basis of a new artificial intelligence system. The AI proceeds to create a sort of utopia in a small desert town while running afoul of an anti-technology group. Transcendence posits, in line with current real-world thinking, that something like a technological singularity will occur at some future point, at which time the exponential growth in self-perpetuating artificial intelligence will leave our world unrecognizable…whether for better or worse.

Star Trek: Discovery (2017 — )

In Discovery’s second season, the crew (under soon-to-be-spun-off Captain Christopher Pike, played by Anson Mount) discovers the AI that guides Starfleet’s intelligence operations, Control, has been going off the books, in part by making a meat puppet of a very dead intelligence chief. (Maybe they shouldn’t have named it “Control?”) A time traveller (The Wire’s Sonja Sohn) warns the Disco gang that, if given access to data that their ship has acquired, Control will eventually wipe out nearly all sentient life, a revelation that launches the ship and the series into its new far-future status quo. As a bonus, the phenomenal Michelle Yeoh buys the gang some time to escape by beating the crap out of the reanimated villain, but that’s probably not a scenario that’s going to benefit you as you contemplate the benefits and dangers of AI.

Malevolence Level: Given Control’s use of reanimated corpses as part of a plan to wipe out all life everywhere, I’d say a solid 10/10 is fair.

Her (2013)

In this Spike Jonze film, Joaquin Phoenix plays a deeply introverted man whose job involves writing personal letters for people who struggle with that sort of thing. In the middle of a divorce, he meets Samantha, a new virtual assistant (a slightly more novel concept in 2013) that he customises with the voice of Scarlett Johansson (taking the job, no doubt, from a real AI). Jonze employs the sci-fi lite scenario to explore the various complications and permutations of modern relationships (an increasing number of which occur online), but also nod to the problem of falling in love with something only designed to seem human, but which has very different goals and motivations. As real-world AI grows increasingly sophisticated, we’ll eventually have to deal with the consequences (for better or worse) of developing attachments to our near-future Siris and Alexas.

Malevolence Level: While offering up some disturbing implications, Samantha is never overtly threatening. (This is one of the few explorations of AI in pop culture that doesn’t at least hint at the end of civilisation.) 2/10.

Omniscient (2020 — )

In near-future Brazil (from whence the series hails), crime is almost nonexistent. Tiny autonomous drones follow everyone pretty much all the time, and any wrongdoing is quickly documented and prosecuted by AI systems that determine exactly what crime was committed and pass sentence. At least until series lead Nina Peixoto (Carla Salle) comes home to find that her father has been murdered without anyone’s drone setting off an alarm. There’s no one to investigate, since crime is virtually nonexistent, and the company’s privacy protocols are such that it’s all but impossible for Nina to gain access to her father’s drone feed. The show considers questions the worth of sacrificing privacy for security in a world where everything is outsourced to machines.

Malevolence Level: Debatable, but the idea of drones following us around and documenting everything that’s not already being documented is creepy enough. 3/10.

2001: A Space Odyssey (1968)

An obvious choice, perhaps, but an essential one. Without ever seeming exclusively or even primarily villainous, the HAL 9000 supercomputer (memorably personified as a single red camera lens) murders several people in its efforts to ensure the success of the Discovery One’s mission. Even given all of that, HAL’s gentle, childlike demeanour makes it hard not to feel a tinge of sympathy when it’s “killed” by Keir Dullea’s Dave Bowman. Like pretty much everything else in Kubrick’s film, there are many valid ways to interpret HAL’s role. I see it as a way to examine the limits of logic and the dangers of cold, inflexible reasoning in the face of human concerns. HAL also reminds us that an AI made in the image of humanity is only ever going to be as good as its programming.

Malevolence Level: HAL’s gentle personality evokes sympathy, but also makes the computer’s cold-blooded actions even more disturbing. 6/10.

Black Mirror: “Black Museum” (2017)

Most Black Mirror episodes turn on the idea that we’re all villains, and that the danger of technology isn’t in the tech itself, but in the ways we’ll use it for our own selfish purposes. This episode, an anthology-within-an-anthology, is unique in that it offers up an actual villain: Rolo Haynes (Douglas Hodge), who runs the museum of the title. The vignettes, presented to museum visitor Nish (Letitia Wright), culminate in a visit to an exhibit all about Clayton Leigh (Babs Olusanmokun), a man who’d been convicted (over plenty of conflicting evidence) of murder. On death row, he was approached by Rolo, who’d made a complete copy of his consciousness before execution. Now Clayton (or, rather, a fully feeling and sentient copy) sits in the electric chair forever, while paying visitors have the chance to punish the (possibly guilty) man over and over again. He represents a different type of artificial intelligence from the strictly machine-born varieties in other stories, but the Clayton AI challenges us to wonder about our responsibility toward the increasingly human-like intelligences of the future and what happens to our already dodgy morality when we can create human-like avatars on whom we can enact our most sadistic fantasies without consequence.

Malevolence Level: The very human Rolo Haynes is the most disturbing figure here; AI Clayton is almost certainly the victim. 0/10.