Who is in control of artificial intelligence and can we trust them?

Abstract

Many films follow rules. Contained within the 120 minutes, or thereabouts, is a sometimes formulaic production led by the rules pertaining to its genre. But some films deliberately flout the norm to challenge the cinematic process and explore big ideas. Whichever boundaries could be crossed, Stanley Kubrick crossed them in the 1968 epic 2001: A Space Odyssey. The project started with a letter from Kubrick to author and collaborator Arthur C Clarke, with an invitation to make the proverbial “really good” science fiction movie. It was expensive (around 3 million US$) and took years to complete, but it is still considered one of the greatest films ever made. The special effects, before computer generated imagery changed the cinematic experience, are phenomenal, and contribute to an experimental approach to structure, storyline, and timeline, where a disjointed sequence of images leave the audience to interpret and analyse with subjective curiosity. So is the end result rigmarole or is it genius?Watch the film and you will likely wonder what it all means—it is our nature to seek relatability, to fit into our parameters of understanding the world. Opening in prehistoric Africa (after a very long musical prelude from Johann Strauss's Blue Danube waltz and an extra-terrestrial sunrise), the coming of man, the discovery of tools and weapons, appear to symbolise the inevitability of violence from power and rule. A shiny large monolith is found by the apes and is then the subject of space missions millions of years later. Two journeys, only 18 months apart, follow and the assumption (from the title of the film) is that we are now in the millennium era. Then, after the failed mission, a glimpse into the infinite future is an alienating (and alien) one, and not one we can yet recognise.London Mathematical Laboratory (LML), partners of the Science on Screen season at London arts venue, the Barbican, recently screened this film. LML, an institute for basic science, encourages its members to “follow their curiosity”, and explore direct and indirect connections between art, philosophy, and science. Professor of Computer Technology and LML fellow Peter Robinson (University of Cambridge, UK) has a particular interest in emotionally intelligent interfaces—computers that understand social signals—and introduced the film with a curiosity of his own. As the boundary between people and computers continually shifts in line with technological progress, against the landscape of Kubrick's opening scene, Robinson asked if the trajectory of evolution might transform humans into robots.What does it mean to be human? From the ancient Greeks to the present day, philosophers, scientists, and artists have explored and theorised on what unique physical, social, biological, and emotional traits distinguish humans from other animals and from computers. Most obvious difference is the brain, with about 86 billion neurons, of which the cerebral cortex comprises 16 billion, comparative to the chimpanzee cerebral cortex of 6·2 billion neurons. And within the physical brain is the mind, where thoughts, feelings, and consciousness exist. Humans have the power of forethought—to imagine the future and create it, and they have an awareness of mortality, beyond a basic survival instinct. Charles Darwin wrote in The Expression of Emotions in Man and Animals that “blushing is the most peculiar and the most human of all expressions”. Could a computer ever simulate the involuntary dilation of the capillaries in human cheeks as a response to feeling embarrassment? The ultimate question arises: can robots ever become sentient beings?Research into the potential of artificial intelligence was in its infancy when Kubrick created HAL 9000 (an artificial intelligence character in the film), but by the late 1960s computer scientists were already working on developing machine learning in robots. Douglas Rain, the voice of HAL, has secured his place in cinematic history as the most influential character in Kubrick's film. “I’m sorry Dave, I am afraid I cannot do that”. Those ominous words signifying the start of HAL's malfunction are not so strange to a 2019 audience. We hear this voice—robotic but “intelligent”, courteous but slightly sinister, in many personal assistant-like machines today.HAL is the Discovery One crew's assistant, but he also controls the aircraft. He has been programmed to withhold the real objective of the mission from the astronauts, and the inference is that his breakdown is somehow related to this deception. Is this human emotion or a technical overload? What is true is that the audience can empathise with him—his character overshadows that of astronaut David Bowman (played by Keir Dullea)—in fact, the humans in Kubrick's future are not particularly alluring; they appear to be lacking in richness, complexity, and personality. HAL is more interesting, and even more expressive. As he gets shut down by Bowman, there is a sense of injustice and loss as HAL almost pleads for his life. But it is only a computer?Traditionally, fictional robots are created with malicious intent; but really they are made in our own image, because evil is not intrinsic in technology. Currently, machines can express emotions, but this is not the same as experiencing emotions. Ethical considerations are focussed on the motivation and competence of the person who makes them. As Robinson says, wisely, “the question is, are we worried about HAL being sentient, or worried about the motivation and competence of the people who made him?” And the film—it either appeals, or it doesn’t. Really it just depends what you want from a movie.2001: A Space Odyssey 1968, directed by Stanley Kubrick Running time 2h 29m Many films follow rules. Contained within the 120 minutes, or thereabouts, is a sometimes formulaic production led by the rules pertaining to its genre. But some films deliberately flout the norm to challenge the cinematic process and explore big ideas. Whichever boundaries could be crossed, Stanley Kubrick crossed them in the 1968 epic 2001: A Space Odyssey. The project started with a letter from Kubrick to author and collaborator Arthur C Clarke, with an invitation to make the proverbial “really good” science fiction movie. It was expensive (around 3 million US$) and took years to complete, but it is still considered one of the greatest films ever made. The special effects, before computer generated imagery changed the cinematic experience, are phenomenal, and contribute to an experimental approach to structure, storyline, and timeline, where a disjointed sequence of images leave the audience to interpret and analyse with subjective curiosity. So is the end result rigmarole or is it genius? Watch the film and you will likely wonder what it all means—it is our nature to seek relatability, to fit into our parameters of understanding the world. Opening in prehistoric Africa (after a very long musical prelude from Johann Strauss's Blue Danube waltz and an extra-terrestrial sunrise), the coming of man, the discovery of tools and weapons, appear to symbolise the inevitability of violence from power and rule. A shiny large monolith is found by the apes and is then the subject of space missions millions of years later. Two journeys, only 18 months apart, follow and the assumption (from the title of the film) is that we are now in the millennium era. Then, after the failed mission, a glimpse into the infinite future is an alienating (and alien) one, and not one we can yet recognise. London Mathematical Laboratory (LML), partners of the Science on Screen season at London arts venue, the Barbican, recently screened this film. LML, an institute for basic science, encourages its members to “follow their curiosity”, and explore direct and indirect connections between art, philosophy, and science. Professor of Computer Technology and LML fellow Peter Robinson (University of Cambridge, UK) has a particular interest in emotionally intelligent interfaces—computers that understand social signals—and introduced the film with a curiosity of his own. As the boundary between people and computers continually shifts in line with technological progress, against the landscape of Kubrick's opening scene, Robinson asked if the trajectory of evolution might transform humans into robots. What does it mean to be human? From the ancient Greeks to the present day, philosophers, scientists, and artists have explored and theorised on what unique physical, social, biological, and emotional traits distinguish humans from other animals and from computers. Most obvious difference is the brain, with about 86 billion neurons, of which the cerebral cortex comprises 16 billion, comparative to the chimpanzee cerebral cortex of 6·2 billion neurons. And within the physical brain is the mind, where thoughts, feelings, and consciousness exist. Humans have the power of forethought—to imagine the future and create it, and they have an awareness of mortality, beyond a basic survival instinct. Charles Darwin wrote in The Expression of Emotions in Man and Animals that “blushing is the most peculiar and the most human of all expressions”. Could a computer ever simulate the involuntary dilation of the capillaries in human cheeks as a response to feeling embarrassment? The ultimate question arises: can robots ever become sentient beings? Research into the potential of artificial intelligence was in its infancy when Kubrick created HAL 9000 (an artificial intelligence character in the film), but by the late 1960s computer scientists were already working on developing machine learning in robots. Douglas Rain, the voice of HAL, has secured his place in cinematic history as the most influential character in Kubrick's film. “I’m sorry Dave, I am afraid I cannot do that”. Those ominous words signifying the start of HAL's malfunction are not so strange to a 2019 audience. We hear this voice—robotic but “intelligent”, courteous but slightly sinister, in many personal assistant-like machines today. HAL is the Discovery One crew's assistant, but he also controls the aircraft. He has been programmed to withhold the real objective of the mission from the astronauts, and the inference is that his breakdown is somehow related to this deception. Is this human emotion or a technical overload? What is true is that the audience can empathise with him—his character overshadows that of astronaut David Bowman (played by Keir Dullea)—in fact, the humans in Kubrick's future are not particularly alluring; they appear to be lacking in richness, complexity, and personality. HAL is more interesting, and even more expressive. As he gets shut down by Bowman, there is a sense of injustice and loss as HAL almost pleads for his life. But it is only a computer? Traditionally, fictional robots are created with malicious intent; but really they are made in our own image, because evil is not intrinsic in technology. Currently, machines can express emotions, but this is not the same as experiencing emotions. Ethical considerations are focussed on the motivation and competence of the person who makes them. As Robinson says, wisely, “the question is, are we worried about HAL being sentient, or worried about the motivation and competence of the people who made him?” And the film—it either appeals, or it doesn’t. Really it just depends what you want from a movie. 2001: A Space Odyssey 1968, directed by Stanley Kubrick Running time 2h 29m 2001: A Space Odyssey 1968, directed by Stanley Kubrick Running time 2h 29m 2001: A Space Odyssey 1968, directed by Stanley Kubrick Running time 2h 29m