Why Does Film and Television Sci-Fi Tend to Portray Machines as Being Human?

This paper identifies, and attempts to explain, a lack of diversity in the way that cinema and television science fiction represents robotics and artificial intelligence (AI). Through a qualitative content analysis of recent film and television portrayals, it is argued, that a limited and limiting vision predominates. This limitation may serve to ideologically reinforce the power of corporate elites. It may also hamper discussion and debate around technological possibilities and their relationship with society. There has been a slew of entertainment productions since 2013 that represent AI and robotics. This work examines Her (2013), Transcendence (2014), Interstellar (2014), Chappie (2015), Ex Machina (2015) and Humans (2015). AI is an old theme in science fiction. What is new, however, is that representations of AI now take place in a world where advanced machine learning is a reality. Nevertheless, rather than exploring existing AI or its future possibilities, popular fiction has remained rooted in a ‘human as machine’ metaphor and the fantastical vision of AI as a form of artificial humanity. The paper argues that fictional representations form part of how we learn about and understand such emerging technologies. They are part of how we think about and discuss the relationship between technology and society today. Science fiction does not just offer speculative representations of social reality it may, in various ways, help to shape it. Science fiction may inspire and inform the work of technologists. It may also form part of the social construction of new technologies, shaping how people see, understand and interact with them. Currently, grandiose and often apocalyptic visions of AI may blind us to the prosaic but serious risks that such technologies can present. Artificial intelligence exists today but in a form quite different to the strong AI that is typically presented in popular culture. It exists in the machine learning technologies that permit, for example, dynamic speech recognition, face recognition, emerging medical diagnostic techniques and so on. Machine learning can be used to gather, integrate and analyse immense data sets for use in sales, insurance, surveillance and so on. If AI poses a threat to society, diminished privacy and the mass unemployment of professionals may be a better place to concentrate rather than dreading the rise of self-aware robotic overlords. Popular science fiction representations are limited for a number of reasons. First, a cybertotalist culture, which flourishes amid the west’s computer industry, propagates a belief in AI as the path to a post-human Singularity (see Lanier 2000). This describes an apocalyptic moment when technological evolution will outrun human control and outstrip humanity’s physical and mental capacities. Such mythology mystifies and symbolically elevates the role of computing. It serves industry by creating an aura around its products. Secondly, we interact every day with machines that pretend to be alive. Think of a computer second guessing your spelling and formatting choices, or Siri saying ‘Sorry I didn’t get that’. Culturally, this predisposes us to visions of computer personhood (see Lanier 2011: 4). Finally, mass market film and television tends to offer a limited vision of AI that, intentionally or unintentionally, resonates with the central tenets of cybernetic totalism. Essentially, human consciousness and computer technology are portrayed as being, in principle, the same thing. Rather than being a product of ideology, however, this type of representation may be due to a central limitation in film and television. Both need to attract large, diverse audiences through identification with characters using universal human traits. Thus, screen science fiction is more likely to portray AI via humanised central characters. Thus machines are predominantly portrayed as being gendered, sexualised and emotional ‘persons’ because audiences are less likely to identify with, or care about, inhuman machines. It is ironic that productions that set out to explore AI, constrain its possibilities by only representing it in terms of its similarity to humanity. This can be understood in light of cinema and television’s need to engage audiences through psychological identification. To follow and enjoy a narrative, audiences need to see some part of themselves or other identifiable types of people in the characters. This simple constraint in making film and television may have consequences for the way that AI is represented and imagined. The idea that creating consciousness in our own image would be the crowning achievement of human science is questionable. It also suggests that there is a kind of species solipsism here. Film and television predominantly imagine machines as versions of ourselves. This may prevent us from seeing AI and robots for what they are, and from more diverse imaginings of what they could become