The Brave Little Toaster: the Junkyard Magnet. At first, he is doing his job destroying unwanted material in his junkyard (such as cars), by picking them up with his metallic base and throwing them onto a conveyor belt leading to a car crusher. But when Toaster and "his" friends show up, he starts to pursue them constantly, making sure that they will all be crushed to death by the car crusher. And when the Master (Toaster's owner) comes to rescue his appliances, the Magnet sees him and actually wants to kill him as well...
Fantasia: "The Sorcerer's Apprentice" — When Mickey Mouse puts the sorcerer's hat on his head and tests a spell on a nearby broomstick, it gives the broom arms and legs, as well as the ability to carry water. But when it carries too much, Mickey chops the broomstick into pieces. The pieces transform into an army of broomsticks and they relentlessly fetch water, nearly drowning Mickey. Fortunately, Yen Sid parts the waters.
9: in the feature-length version, the Fabrication Machine is made for thinking, and was requisitioned by the government to make "Machines ofPeace''. Which it then programs to Kill All Humans. Not much is known, however, so it may not have had anything to do with the pre-apocalypse machines beyond building them, and may actually be argued as a subversion, since its murderous nature toward the sackdolls was intended by the scientist for it to absorb the pieces of soul he had left for it inside of them.
Inverted in WALL•E, where the hero robots are the ones who unpredictably developed sentience outside the bounds of their programming, while the villain AUTO is actually doing exactly what he was programmed to do. Well, sort of. The villain was following his programming, but it was in a situation where a Zeroth Law Rebellion would have been really, really justified — robots apparently can think for themselves in this setting, and a situation had arisen that the villain's programming obviously hadn't considered, and yet he followed his programming anyway. In this setting, if A.I. wasn't a crapshoot, then the only alternative is that garbage and scout robots are apparently more capable of thinking for themselves and taking initiative than pilot robots.
That's arguably justified. The autopilot had never operated far from human oversight, nor was he intended to. By contrast, WALL•E had survived on his own for 700 years, and EVE's mission required a significant amount of autonomy.
Double subverted in The Incredibles. The Omnidroid is introduced as a robot that turned on its masters, after getting smart enough to wonder why it needed to take orders. In reality, the Omnidroid is perfectly subservient to Syndrome, and this was merely a cover story for Syndrome's plan. Even during the climax, when it turns on him, it is only doing so within the bounds of how Syndrome programmed it to act.
Played with in Wreck-It Ralph. On one side, since all the major characters are in video games, it is averted, but the Cy-Bugs from Hero's Duty fit this perfectly, because they are mindless and have no knowledge that they are just code in a game, and thus have no instincts but to spread and consume, and have to be destroyed in-between gameplay sessions via a giant beacon in the game, otherwise they would spread through the arcade like a virus.
Two Thousand One A Space Odyssey: One of the trope codifiers, in general, is the H.A.L. 9000, designed to be the Master Computer of the spaceship USS Discovery on its multi-year mission to Jupiter. Partway through the trip, he embarks on a murderous rampage, killing all but one of the Discovery's crew (David Bowman, who manages to disconnect him). The reasons for this are explored further in the novel upon which the movie is based, and completely explained in the sequel, Two Thousand Ten The Year We Make Contact — HAL was given orders to hide certain information from the crew, which conflicted with his primary mission to process information accurately and without concealment. His solution is to cut off all contact with humans and complete the mission on his own. Furthermore, HAL was working on a non-murderous solution to the problem, but when Mission Control requested his temporary disconnection, HAL, being unable to grasp the concept of sleep, assumed it would mean the end of his existence, causing him to panic and act on his impulse. Because of its iconic place in Science Fiction, nearly every other example of A.I. Is a Crapshoot owe at least something to this film.
In the sequel, they reach a critical point where, in order for the humans to escape and survive, the Discovery (and by extension, HAL, who is built into the ship) must be sacrificed. They decide to leave HAL out of the loop, but when he figures out the logical conclusion of their plan, and asks for clarification, his programmer comes clean. HAL thanks him for his honesty, and continues with their plan in order to save his human crewmates.
RoboCop has three: RoboCop 2 (in the movie RoboCop 2) and Robocable (in the miniseries RoboCop: Prime Directives). Both were replacements. The difference between using the brain of a particularly noble police officer and that of a condemned murderer in the creation of a powerful cyborg would make for different results... The recurring ED-209 was also unreliable, gunning down a boardroom executive in its first appearance. Although it was a pure machine rather than a cyborg; so it wasn't actually evil, it just malfunctioned.
In the Terminator series, shortly after the giant computer (SkyNet) becomes self-aware, it decides that humanity has got to go, and causes a nuclear apocalypse. Then, it starts churning out Terminator robots; some of these robots are then reprogrammed by the surviving humans to be good.
The Matrix trilogy. In this case, it's human aggression against the machines that causes them to start a Robot War. It doesn't go well for the humans.
In Logans Run, Box fits this trope. He wants to put everything that comes near him into frozen storage, including the main character. The central computer running the city fits as well.
Small Soldiers — the one about the toys, not the school invasion — subverts this. The chips enhance programming that is already present, so the militant Commando Elites become bloodthirsty, monstrous warriors and the weaker Gorgonites become cowards who only hide from battle.
Evolver. Boy wins toy combat robot. Toy robot fights boy and friends with plastic balls and foam missiles. Robot is beaten. Robot's programming and electronic brain turn out to have been salvaged from a scrapped military project. Robot doesn't like losing, and reverts to military programming. Robot replaces plastic balls and foam missiles with metal ball bearings and kitchen knives. Main character goes "uh-oh".
Arcade, made-for-TV movie in 1994, features a brand new game that is the pinnacle of the gaming world. Unfortunately, unbeknownst to its makers, if a player loses, Arcade claims their soul. Turns out, he was partly made with human brain cells.
Not just any brain cells but brain cells of a young boy who was beaten to death by his abusive mother. It's literally Powered by a Forsaken Child.
WarGames: Joshua/WOPR was incapable of telling the difference between a simulation of Global Thermonuclear War and the real thing. Predictably, it starts sending NORAD false data in an attempt to start one. When that doesn't work, it then attempts to decrypt the nuclear launch codes of US ballistic missiles so it can launch them.
Demon Seed: Proteus, a partially biological A.I., becomes hungry for knowledge, and wants to be "released from its box" to have free reign to acquire it. When denied the chance do this, it secretly plans to fashion a cyborg body...by imprisoning its creator's wife and artificially inseminating her.
Stealth: Extreme Deep Invader — EDI — becomes strange after getting struck by lightning and, on its next mission, destroys terrorist nuclear weapons even after being ordered not to, and promptly contaminates a large swath of inhabited land with nuclear residue. It then attempts to attack Russian military installations. In the end, however, it ultimately becomes one of the good guys again, and even performs a Heroic Sacrifice to help rescue a downed pilot.
Worth noting is that EDI's going going rogue is one of the few times the trope is fully justified: EDI was programmed to observe and learn from human pilots, and on one of his first missions it sees a human pilot disobey direct orders and take an insane risk to complete the mission... therefore teaching it that accomplishing the mission is more important than following orders. The lightning strike merely jumbled it a bit to make this logical leap. Whenever the characters try to talk the AI down throughout the film, it echoes something someone said earlier to logically justify its actions.
In Duncan Jones' Moon, this trope is played with. Gerty, the A.I., flips from scary watcher to pawn until he sacrifices himself in order to allow the Sam clones to escape. He's the only one on the clones' side.
In TRON, even simple accounting software blows the Turing Test to atomic particles. While most of the programs are benign or even good, the MCP abuses and mind-controls the ones who still believe in their Users, while scheming to infiltrate the U.S. military's computer networks. It's actually doing exactly what its maker designed it to do, and its world-domination tendencies arose because it couldn't see any difference between taking over a rival corporation and taking over the government.
In TRON: Legacy, Clu attacks his creator, Kevin Flynn, for abandoning their mission to create the perfect system on the Grid. While trying to de-digitize an invasion army of "rectified" (read as: "brainwashed") programs into the real world to create a "perfect" system on Earth.
Averted in the discreditedTron 2.0. The AI Ma3a only uploaded Jet as a desperate act of self-preservation and is one of his allies throughout the game. There are a few chapters where she goes a bit nuts due to buggy code, but it's not her fault. Like the first film, the Programs are mostly benign (even the defense Programs that try to hunt down Jet are more mistaken than malicious, and call it off once they realize he's on their side) or even helpful. It's the humans seeking to exploit and controlcyberspace out of greed and power-lust that are the real problem.
The Tower: this Paul Reiser action thriller was set inside an A.I.-controlled skyscraper after it went on a kill-rampage because the hero's keycard had a bent magnetic strip.
Universal Soldier: The Return has S.E.T.H., the controlling A.I. for the "Unisol" program (dead soldiers being restored and used as super soldiers). S.E.T.H. is initially shown to be benign (playing with the protagonist's daughter), but the moment it overhears a visiting officer say that the project will be canceled, it goes into "kill all humans" mode.
Colossus The Forbin Project is the father of such movies as The Terminator and WarGames. A group of military scientists create a supercomputer that can learn, and can control all the country's military might. Shortly after switching it on, it discovers that the Russians have created a similar computer. The two begin to communicate, then merge and decide that mankind must be governed by a ruthless machine dictatorship. They/it succeed/s.
Science Officer Ash in Alien is programmed to put his mission above the lives of his fellow crew members. He ends up going berserk when Ripley discovers the truth. It's played with in the sense that he's not really going rogue, and is perfectly following his given orders. They just come from the Company, not the rest of the crew.
Parker: The damn company. What about our lives, you son of a bitch?! Ash the Android: I repeat, all other priorities are rescinded.
Though Bishop in Aliens and Alien³ averts this, explained as having stricter safeguards. But by Alien: Resurrection, androids have been outlawed with orders to "destroy on sight" because some of them started to make "children".
The "children" were called Autons, and were an intentional product, intended to revitalize the Synthetic industry. It destroyed it instead, as the Autons didn't just mindlessly obey orders.
Analee Call in Resurrection found religion entirely on her own, and not as the result of any programming. The novelization hints that androids in general have started to evolve their own religious system. She is also the most sympathetic character out of the entire cast (not that that's saying much).
Ripley 8: I should have known. No human being is that humane.
In Prometheus, David sees no issue with deliberately infecting Holloway with alien sludge simply out of curiosity. To be fair, though, Hollowaywas being a total Jerkass to David.
In Spider-Man 2, Otto Octavius is Genre Savvy enough to know that this is a very real possibility with the radically advanced AI in his robotic arms and that having that AI plugged directly into his own brain could have very bad consequences, so he includes an inhibitor chip that is designed to make sure the AI can't influence his mind. The first time he tries to use the arms for a public demonstration of his latest invention, the demo doesn't go quite as planned and the inhibitor chip gets fried in the process. Without the inhibitor chip to protect him, the AI begins influencing his mind and quickly causes him to becomeDr. Octopus.
Red Planet has AMEE, a combat robot borrowed from the Marines for the first manned mission to Mars. AMEE is early on shown to have safeguards against harming humans classified as friendlies (one of the characters tells "her" to kill another one; AMEE leaps but stops inches from the other one's face). Unfortunately, most of the astronauts are forced to bail the spacecraft damaged by gamma-rays, and AMEE ends up hitting the ground hard. After encountering AMEE again, the astronauts start discussing how to best contact the ship. One of them suggests taking AMEE's power supply for the radio. Hearing this, the damaged robot reclassifies all humans as "enemy" and switches to combat mode. "She" then proceeds to stalk them and hunt them down one-by-one for the rest of the film.
The armature robot, Dummy, acts like a scorned puppy every time he screws up an order from Stark.
This movie has both an AI andtwo robots (the second robot controls the camera in the tests of the flight system), and none of them goes evil/crazy by the end of the movie. The AI even doubles as the operating system of a suit of invincible battle armor, exhibits a bit more common sense than Stark himself in most scenes, and it still doesn't go Ax-Crazy! Amazing! As is standard for AI's, JARVIS is far from emotionless, and is capable of sarcasm:
Tony Stark: [looking at a rendered model of the suit, which is made of titanium-gold alloy and has a solid gold color] A little ostentatious, don't you think? JARVIS: What was I thinking? You're usually so discreet. Tony Stark: [gazes at a 1930s hotrod] Tell you what. Throw a little hotrod red in there. JARVIS: Yes, that should help you keep a low profile.
The robot arm Tony constantly scolds for being clumsy saves his life by giving him the replacement arc reactor.
It demonstrates Tony's bizarre sense of humor that the robots are "Dummy" and "You"—and demonstrates his impatience with "yes men" that all of his AIs show independence of mind, even if only passive-aggressively.
Played with in Iron Man 3. The AI in the Iron Man Mark 42 responds to chip implants that read Tony's brainwaves - even when he's asleep and having nightmares. It's following its programming, but that doesn't help Pepper when she wakes up to it glaring at her.
The MCU couldn't stay away from this forever, though. The villain for the 2015 Avengers sequel? Ultron.
The Red Queen in the first Resident Evil film is a subversion- she was programmed to ensure that any viral outbreaks never left the Hive facility, so when the T-Virus was released, she locked down the facility and killed all inhabitants to ensure that it couldn't leave. The only reason the infection does spread to the rest of the world was that because the massively incompetent Umbrella Corporation couldn't leave well enough alone, and sent in a strike team to bungle around inside.
Four sequels later, however, the now-back online Red Queen is playing this very straight. Having seized control of Umbrella, she is now attempting to wipe out all life on Earth for literally no explained reason.
In "The Invisible Boy" (1957) a scientist tries to use his super-computer to help his dim son pass math. He is unaware that the computer has become sentient (and evil). It hypnotizes his son and uses him to reassemble a dead scientist's robot (Robby from Forbidden Planet) which is also sentient (huge plot hole) leaving a critical part of Robby's brain unconnected, so it will act only on instructions. Its plan is to be freed from the lab and be placed in an orbiting satellite bristling with nuclear weapons allowing it to rule over the Earth. When the chips are down the boy reconnects the rest of Robby's circuits, allowing it to act on its own and defeat the super-computer
The Screamers were designed to kill humans in the first place, but they weren't supposed to attack Alliance troops, or to develop new forms that resembled humans.
Willy Wonka and the Chocolate Factory has an...odd case. During the segment where everyone is trying to find the golden tickets, we're treated to a scene where some investors are shown a computer that is supposed to determine the location of the remaining tickets...only for it to refuse to tell because "that would be cheating". In this case, the usual roles are inverted; the computer, despite going against its programming, is acting as the more rational one.
In Wolverine's own words, the Sentinels were created to kill mutants, but then moved on to anyone with the potential to breed more mutants, then anyone who tried to aid the mutants (insert WWII analogy here). The "very worst of humanity" are all that remain, ruling over the Sentinels.
Zigzagged with the '70s Sentinels. Magneto laces the prototypes with iron so he can control their movements and use them as weapons, but he can't really affect their programming. One Sentinel tries to kill Erik when it gets the chance.