Before reading this article, please click "Follow" to facilitate your discussion and sharing, bring you a different sense of participation, thank you for your support
Science fiction movies play the role of early warning of human civilization, and one of the eternal themes is to ask the significance of scientific and technological development for the future of mankind.
One of the forms of expression of science fiction films is to reflect on the interrelationship between technology and society, technology and civilization, and technology and human nature while fearing that man-made products in turn rule or harm themselves.
The test method for testing the robot's consciousness in the film is not actually the Turing test, but it has effectiveness and superiority, and its essence is the distinguishing standard of "human-machine consciousness", and the criterion for judging whether the robot has consciousness lies in "understanding".
The criterion for distinguishing between machine intelligence and human intelligence lies in "feeling", although intelligent robots "can" understand human ethical principles but do not necessarily agree, so robots will not have different ethical principles than humans.
Next, let Xiaoyu take you to understand how "Ex Machina" passed the "Turing test" to determine whether the robot has a "human heart"?
Synopsis of the story
Ex Machina, a small-budget sci-fi film directed by Alex Garland, first released in 2014, tells the story of how a consciously intelligent robot uses humans to gain freedom.
Programmer Gary Kaleb was invited by the company's boss Nathan to go to his private villa for a week-long "vacation", but in fact, it was Nathan's premeditated arrangement, and Kaleb was invited to conduct a "Turing test" on the robot he developed, "Ava".
During this time, Ava uses her wisdom to make Kaleb fall in love with herself and promise to help her get free.
At the end of the film, with the help of Kaleb and another female robot, Kyoko, Ava kills Nathan, who created and imprisoned her, but also abandons Kaleb who helped her, and finally, Ava walks out of the villa alone and enters human society and is free.
The style of the entire film of "Ex Machina" from low-brightness tones, to regular sets, to simple dialogue design, all carry a rational coldness and have a sense of technological abstraction.
The whole film implements the principle of "minimalism", there are less than five characters in the whole film, the main plot is limited to a villa, and it is regularly divided into seven chapters, corresponding to the seven days spent by the male protagonist.
The film provides rich and vivid elements for thinking about the future of artificial intelligence technology, and the theme of "artificial intelligence robots" discussed is related to people's imagination and worries about AI in reality, and the Turing test method is carried out on the theme of "human-machine relationship":
Will AI robots be self-aware? They will all pass a robotic test to get an answer...
Robot testing
In "Ex Machina", due to the reversal of the plot, the real robot test method is not revealed until the end.
At first, Kaleb thought he had won the company grand prize - he could go to the boss's villa to spend a vacation with the boss, but upon arrival, he was invited to participate in the so-called "Turing test" and play the role of a human in the "Turing test".
So, does the method used to test the robot's consciousness in the film belong to the category of "Turing test"?
In 1950, Turing proposed in the article "Computer and Intelligence" the standard for judging whether a computer has human intelligence, that is, a person and a computer are put behind the scenes, and testers are asked to judge which is a computer.
If the judgment is wrong, it is considered that the computer has passed the Turing test and has human intelligence, in this sense, the test method presented at the beginning of the film is not the Turing test.
Because in "Ex Machina", Kaleb can "see" Ava directly, and the circuits and other mechanical parts on Ava's body clearly indicate her robot identity.
In the traditional Turing test, the robot is behind the scenes, and only in this way can the human tester distinguish whether the object with whom it communicates is a robot is credible.
But in a broad sense, if you only regard the Turing test as a method of "judging whether a machine is intelligent through human-computer dialogue", the test mode in the film is somewhat different, but still belongs to the category of Turing test.
Stephen Hanard has pointed out that the narrow Turing test only tests the verbal intelligence of robots, and he believes that only through the more "complete Turing test" can be passed.
That is, robots are difficult to distinguish from humans in the dual abilities of symbolic ability and behavioral ability, so as to avoid the questioning of Searle's "Chinese argument".
However, the end of the story ushered in a reversal of the plot, it turned out that Kaleb did not play a human in the Turing test, but a tool used by Ava to fight for freedom.
Nathan compares Ava to a rat in a maze, and the only way out he points out to Ava is "Kaleb".
In order to escape, Ava must make good use of him, using everything such as self-awareness, imagination, means and strategy, and femininity to successfully arouse Kaleib's sympathy, so as to achieve her own goal of escape.
And this is the real robot test content in "Ex Machina", and the result of the test is obviously that Ava did it, she passed the test, and Kaleb finally betrayed Nathan in order to protect her.
It can be seen that the real tester here is actually the boss Nathan - a person who has almost no interaction between Ava and Kaleb.
The real test question is not "Does Kaleb think Ava is conscious", but "Can Ava consciously use all her successes to deceive Kaleb in order to fight for freedom".
Therefore, the real robot test in the film "Ex Machina" does not actually belong to the category of the Turing test.
Before the real test method was revealed, Nathan explained to Caleb the question of why the modified Turing test was:
"If you don't see Ava and only hear sound, she will pass the human test, and the real test is to let you see that she is a robot, and then whether you still feel that she is conscious."
In other words, Nathan is confident that his artificial intelligence robot can pass the traditional Turing test, so here he upgraded the difficulty and asked for a more comprehensive intelligence test of Ava.
In Ex Machina, when Ava meets Kaleb for the first time, Kaleb is completely shocked by the smooth conversation between the two.
When Ava was asked "when she would learn to speak", she replied: "I always knew how to speak... This is strange because language is supposed to be learned. ”
This means that if tested only by dialogue according to the standards of the traditional Turing test, Ava's incredible language ability will allow it to successfully fool the test taker in dialogue with humans.
However, such an improved Turing test is hardly the optimal version or the most effective method of the Turing test, and it will inevitably be questioned by cognitive science scholars, thus affecting the science of "Ex Machina" as a science fiction movie.
So, director Alex Garland cleverly uses the reversal, and the real test is not the Turing test, but a "lie" test, that is, whether Ava will lie and know the purpose of lying.
And this test is actually based on the success of the Turing test, because Kaleb has indeed regarded Ava as a "human" and will fight Nathan for her freedom.
The main reason for adopting this real testing method in the film is, on the one hand, the need for plot reversal, or the director's clever arrangement of the plot.
For the protagonist Nathan, he needs to see the most objective answer, so he deceives Kaleb in advance to ensure that the real test can go smoothly.
Therefore, even if there is a reversal, it is logically reasonable from the story, on the other hand, it is also the main reason, which lies in the criteria for judging whether the robot is conscious.
Are the methods in the film effective in testing whether the robot is conscious? If not, is there a more efficient way? If so, which method is better than other test methods?
So, if you compare the method in the film with other methods for testing the robot's consciousness, which method is better? One of the most important objects of comparison must be the "Turing test".
As mentioned above, one of the questioned aspects of the Turing test is that the intelligence it detects is only one aspect of intelligent behavior, that is, verbal intelligence.
And Hanard's complete Turing test is based on verbal intelligence, and expands the scope of other intelligence that needs to be detected.
However, this point is not the most fatal point of the "Turing test", and the two questions that pose the greatest threat to the Turing test method are the "Chinese house" experiment and the problem of his heart.
So, what exactly is the "Turing test" in the movie?
Turing test
In 1980, Searle devised a thought experiment in which a man who knew only English and did not understand Chinese was locked in a room, leaving him with only a small window and a guide book translated from Chinese to English.
At this time, the person outside the house handed him some pieces of paper with characters through the window, which he could translate according to the instruction book in his hand, and wrote Chinese note in the same way and handed it out the window.
For people outside the house who know how to Chinese, the questions he asks Chinese are properly answered by the people inside, and he will think that the people inside the house actually understand Chinese, but this is not the case.
All he did was "mimicry" the pattern matching according to the instructions, and similarly, a computer could not understand the information being processed even if it could pass the Turing test with a function.
Searle thus concluded that "the program itself cannot constitute the mind, and the formal syntax of the program itself cannot ensure the emergence of mental content" in order to oppose the symbolist claim of strong artificial intelligence.
The questioning of the "Chinese house" argument means that if the Turing test is used, people will not be able to distinguish whether the intelligent behavior displayed by the machine is based on highly similar "imitation" or real "understanding", and thus cannot distinguish whether the machine really has intelligence or consciousness.
In the movie "Ex Machina", although the boss Nathan still judges whether she is intelligent by observing the behavior shown by Ava.
But he sets up a relatively complex scenario for Ava to play freely, thus proving that Ava's behavior is governed by "her" mind.
This also proves that Ava has consciousness or intelligence, rather than directly using the method of the Turing test, that is, the test person makes a judgment.
It can be seen that the film's efforts in answering the question of "his heart" are superior to the Turing test.
From the above discussion of the human-machine relationship, it can be seen that the real robot test in the film "Ex Machina" does not actually belong to the category of the Turing test.
The reason why the film chose this non-Turing test method, on the one hand, is due to the need for plot reversal, on the other hand, and the most critical reason is that only with this method can the robot really detect whether it is conscious.
The reason why the test in the film can only be completed in this way is because, neither the traditional Turing test nor the improved Turing test can answer the "Chinese House Experiment" proposal:
The question of how to tell whether the intelligent behavior exhibited by a machine is based on a highly similar "imitation" or a real "understanding" cannot answer the "other mind" question.
The test method in the film can solve these two difficult problems, or at least provide a clear solution to these two problems.
"Ex Machina" uses the "comprehension" criterion to identify whether the robot is conscious, separates the comprehension criterion from the sensibility standard, and re-distinguishes the boundary between humans and robots by analyzing the difference between human-machine "feeling" and ethical principles.