This trope is an often complex puzzle, trap, or maze — either physical, logical, or psychological — that a robot, android, or similarly synthetic but sentient or near-sentient being must navigate in order to be deemed "conscious" or self-aware.
The original Turing Test thought experiment was initially devised to see if a robot or similar non-human could pass as a person by its ability to, essentially, carry on a text chat conversation. If the tester could not reliably distinguish the AI from humans in a blinded test, it passed.
However, both ubiquitous advanced AI chatbot technology and some varieties of the Internet Jerk have recently caused anxiety about the classic Turing Test's inadequacy for determining true consciousness. In the light of this, there has been a proliferation of storylines where engineers attempt to test their life-like creations for signs of real sapience — whatever that might be — with complex psychological games or tests of lateral thinking.
This is a Sub-Trope of the Turing Test trope, which covers methods of determining the sapience of artificial intelligence in general.
Examples:
open/close all folders
Comic Books
- In issue #95 of Legion of Super-Heroes, during the "postboot" Legion's extended stay in the 20th Century, the Legion crosses over with the Metal Men — by attacking them out of the blue to steal the responsometers that grant them sentience. Brainiac 5, who needs advanced technology to build a computer to control a Time Machine, continues to remind his teammates to ignore the objections of the Metal Men as just programming. However, when Platinum and Veridium (Dr. Magnus, who had undergone an Emergency Transformation at the time, along with a Relationship Upgrade with Platinum) sympathize with Ultra Boy and Phantom Girl's mid-battle relationship issues, Brainy immediately orders the Legionnaires to stand down. As he explains to both teams, Brainiac 5 recognized that the two Metal Men had both identified themselves with, and empathized with, the other couple — meaning that they're "living, feeling creatures" who "only a barbarian" would experiment on. (Not only do the Metal Men forgive him, Veridium actually volunteers for Brainy's experiment.)
- One issue of The Transformers: More than Meets the Eye brings up the Ambus Test, a questionnaire for detecting sentience in robots, used by the Transformers, a species of sentient robots. Which fails to recognize sentience in some of their own members, if interpreted strictly. Cybertron has an unpleasant history of Fantastic Racism and Robots Enslaving Robots, so this is presumably a holdover from that time despite continuing to be used into the present.
Films — Live-Action
- Blade Runner:
- Replicants (and one female Replicant in particular) have their sentience questioned by Blade Runner's main character, and the other Blade Runners hunting them, using the "Voight-Kampff" test. This test requires the suspect to answer a series of emotionally provocative questions and scenarios while monitoring their involuntary responses. Ultimately, the movie's central struggle and reveal confirms the sentience of the robots once thought only partly human.
Holden: Describe in single words only the good things that come into your mind about... your mother.
Leon: My mother?
Holden: Yeah.
Leon: Let me tell you about my mother. [pulls out gun and shoots Holden] - Blade Runner 2049 features an updated version of the previous test, where Replicants are continuously berated with even more emotionally manipulative questions by an unseen interrogator in another room. In an odd inversion of the first movie's VK test, where it was used to determine whether humans were acting too robotic (and thus Replicants in disguise), the Baseline Test is used to determine whether Replicants are becoming too humanly empathetic or emotional (and thus unreliable).
Interviewer: Officer K-D-6-dash-3-dot-7, let's begin. [...] Recite your baseline.
K: And blood-black nothingness began to spin... A system of cells interlinked within cells interlinked within cells interlinked within one stem. And dreadfully distinct against the dark, a tall white fountain played.
Interviewer: Cells.
K: Cells.
Interviewer: Have you ever been in an institution? Cells.
K: Cells.
Interviewer:Do they keep you in a cell? Cells.
K: Cells.
Interviewer: When you're not performing your duties do they keep you in a little box? Cells.
K: Cells.
- Replicants (and one female Replicant in particular) have their sentience questioned by Blade Runner's main character, and the other Blade Runners hunting them, using the "Voight-Kampff" test. This test requires the suspect to answer a series of emotionally provocative questions and scenarios while monitoring their involuntary responses. Ultimately, the movie's central struggle and reveal confirms the sentience of the robots once thought only partly human.
- Ex Machina: Normally, a Turing test would be done blind with the tester interacting with both an artificial intelligence and a human, and the AI fails the test if the tester can tell which is which. In Ex Machina, the robot's creator asserts that his robot Ava is beyond this test and could pass it easily. Instead, he wants Ava to convince the tester that she is "human" even though he knows she's a robot.
- I, Robot: Sonny is interrogated and his sentience is assumed nonexistent, but, through a complex rescue plot, Sonny demonstrates his sentience via art and complex emotional communication and understanding.
- In the first Short Circuit movie, Johnny 5's sentience (versus his non-struck-by-lightning brethren) is hotly debated, with the military who created him hell-bent on getting its "property" back and the heroes trying desperately to prove that he is "alive". Ultimately, he is able to prove his sentience through an inkblot test (during which he not only identifies the chemical makeup of the blotting material but is able to recognize the blot's resemblance to other things) and displaying a sense of humor (telling a joke, in response to which he laughs once he gets the punchline).
Literature
- Do Androids Dream of Electric Sheep?, the inspiration for Blade Runner, also features the Voight-Kampff test, which is explicitly designed to detect empathy by measuring involuntary responses. Deckard is only able to identify Rachel as inhuman by noticing that she took a fraction of a second too long to be squicked when he claims his wallet was "pure baby-hide".
- Due to the provable existence of spirits and souls in the setting of No Need for a Core?, and the connection of souls to minds, anything that appears to have a mind is presumed to have a true mind and soul. It is better to err on the side of caution and presume that any appearance of free will represents actual free will.
- Xanth features a metaphysical variant. A manticore travels to Good Magician Humphrey to ask whether or not it has a soul. Humphrey answers that if it didn't have one, it wouldn't be concerned with such things. The manticore is satisfied with this answer because a simple "Yes" or "No" might be a mere guess, whereas this explanation makes the answer self-evident.
Live-Action TV
- Star Trek: The Next Generation:
- In "The Measure of a Man", Data is placed on trial by Starfleet to prove that he's a life form in order to avoid being reassigned to destructive reverse-engineering. Picard manages to win the case by arguing that Data is sentient by the prosecution's own standards, and that the end result of the research, the de facto enslavement of the Data-type androids that would be built without rights, would contradict the Federation's fundamental ideals.
- In "The Quality of Life", a test is created for the exocomps to determine their sentience as follows: 1. Send an exocomp to complete a repair. 2. Simulate a systems failure that would result in damage to the exocomp if it was real. 3. If the exocomp ignores it and fulfills their programming, it's just a machine. If it abandons the job to save itself, it's alive. 4. Recall the exocomp and evaluate the results. The experiment fails (i.e., the exocomp keeps working and is "destroyed"), and is repeated 34 more times just in case. Then a distraction causes Step 4 to be delayed. This time the exocomp wasn't ordered to return, revealing the truth. The exocomp knew the failure was an illusion, so it finished the job and then turned off the alarm caused by the "failure". The exocomp is alive.
- Westworld: Arnold creates the maze as a way of testing the consciousness of the hosts, specifically Dolores and Maeve, both of whom have to suffer through their own deaths and strive to overcome extreme adversity in order to prove (through their own suffering) that they are in fact sentient.
Radio
- John Finnemore's Souvenir Programme: A very odd example in one sketch, where a man suffering the usual difficulty getting through tedious bank security checks turns out to be a robot, who has been lured into a trap to determine he's a robot, for no readily apparent reason. Also, the teller is also a robot, because it takes one to catch one.
Video Games
- CrossCode both references the Turing Test by name and posits an alternative test based on the premise that the AI being tested is presenting themselves as a humanlike AI rather than trying to fool the human tester. This is used by Ivan, a shareholder of the company who developed the MMO that the game takes place in, to test Lea, a sapient AI mentally cloned from a human. The way Ivan performs the test also accommodates the fact that Lea is mostly mute aside from a few words. The test has three basic steps:
- 1. The tester and the subject being tested meet and the tester proceeds to make light conversation, deliberately getting certain details wrong (like the subject's name and the weather), testing not only how the subject makes conversation, but also their ability to react to and correct inconsistencies.
- 2. The tester then proceeds to ask a few simple questions - firstly, what the subject's favorite vegetable/fruit is, a question about a short story the tester tells, and then the color of the vegetable/fruit the subject answered with for the first question. This tests the subject's ability to answer questions that are subjective, involve comprehension of details that lesser AI would trip over, and answering a question based on remembering an answer to a previous question.
- 3. To test if the subject is indeed an AI and not a human being posing as one, the subject is provided a minute-long recording which has a math question, which is processed by the AI rather than listened to normally. The subject must answer the question immediately after processing the information. A human would be incapable of processing that information without listening to it normally or so quickly even with some kind of neural upload, meaning a subject who answers the question correctly immediately after being provided the recording is indeed an AI.
- "Detroit: Become Human" has the Kamski Test. Elijah Kamski, the creator of androids, has Connor choose between shooting another android (Chloe) and getting information or not shooting her and getting nothing to see whether machines really can have empathy.
- Stellaris: The flavor text for the Robotics: Citizen Right law states that all robotics who prove to be self-aware are to be treated as living beings, hinting on the existence of such a test.
- The Talos Principle: Early on, the player character must take a questionnaire to prove he's "really human" in order to get full access on a computer system. This is just a feint, though, as the administration program is a Commander Contrarian that undercuts every argument with slightly circuitous logic, and ultimately suspends the qualifying criteria of being human in order to make a point about the pointlessness of the distinction. The player gradually becomes aware that the character is not a human, but an AI designed to become human-like through directed Mechanical Evolution. It eventually becomes apparent that the entire game world is one vast test to ensure the completion of this goal: the player character solves puzzles in order to prove logic and abstract thinking, but can only escape the simulation by willingly defying a direct command from an established authority, something a mere robot could never do.
- The Excuse Plot for The Turing Test is that the crew left on the Europa base wanted to keep the AI TOM from gaining entrance using a series of puzzle rooms requiring lateral thinking to complete. Since the player is acting as TOM, instead of the human protagonist, the puzzles are ultimately ineffective.
Webcomics
- Freefall: Florence once tested a couple of robots for sentience by asking them "What does your name smell like?"
The non-sentient one simply concluded that names cannot have scents and ended the conversation; the sentient one reasoned that while he had no sense of smell, Florence did, and for all he knew names having scents is a thing among Bowman's Wolves, so therefore the only way to answer the question would be to ask her.
- The Inexplicable Adventures of Bob!: When Princess Voluptua meets Roofus the Robot, she asks him if he's conscious. He says he doesn't know, and she takes this as proof that he is conscious, because a non-conscious machine designed to feign consciousness would just lie and say yes. Conversely, if a robot starts parroting your extreme political views and tries to collect your personal data, it's Just a Machine and fair game to smash up.
- Parodied in xkcd's "Turing Test", in that the computer is so sentient that it somehow manages to convince the examiner that he's the computer all along.
- In "Constructive", one character proposes a new type of Captcha test that will make users rate message board comments as either constructive or not, and then reply with their own comments that will be rated later. The other character asks what will happen if spammers train bots to automatically make constructive and helpful comments? The reply: Mission. Fucking. Accomplished.
- In "Constructive"
