Introduction: AI narratives and science-fictional AI

In 2018, German Chancellor Angela Merkel met the humanoid robot Sophia produced by Hanson Robotics for a conversation event; the year before Sophia became citizen of Saudi Arabia as the first robot being granted the right of citizenship of a country. Another year earlier, in 2016, the software Alpha Go beat the world champion Lee Sedol in the board game Go—20 years after Deep Blue won against the then world champion in chess Gary Kasparov in 1996/1997. The progress in the broad field of artificial intelligence (AI) seems to be catching up with many films of the science-fiction (SF) genre, in which we have been watching humanoid robots and powerful computers for decades. The fictional android Data from the Star Trek franchise, for example, is a valuable member of the Enterprise crew and earned ‘his’ right to personal freedom before a trial; in the same vein, the board computer HAL 9000 from 2001: A Space Odyssey appears to have its own will and by playing chess with the human astronauts “depict[s] the future birth of a superior intelligent being” (Bory 2019: 628)—nowadays, this certainly applies to a software mastering the much more complicated game Go. It is, therefore, not surprising that SF has become a reference point not only in the media, but also for humanities scholars and social scientists on the ethics, opportunities, and risks around AI.

There are two ways to look at SF featuring AI, which I call science-fictional AI: First, it can be viewed as being a substantial part of a larger corpus of AI narratives. Narratives in general are cultural artefacts of various kinds that tell stories, which convey particular points of view or sets of values (Bal 2009). The term AI narratives applies to narratives featuring intelligent machines (The Royal Society 2018: 5), they can be analysed as a reflection of our hopes and fears towards these technologies and thus may shape the development of AI by influencing developers, public acceptance, and policy makers (Cave et al. 2020; Cave and Dihal 2019: 74). In this sense, AI narratives are understood as a serious representation of the potential of real AI and its possible consequences—like foresight or technology assessment. However, science-fictional AI—like the genre of SF in general—is not only about the hopes and fears of the particular technology, but about human dramas for a human audience and readership. From this perspective, it is not AI per se that inspires dramatic stories, but—quite the other way round—the desire to tell dramatic stories requires certain types of AI, for example humanoid robots or almighty systems. Thus, second, science-fictional AI is not necessarily about the technology but can be a metaphor for other social issues.

This second perspective is relevant, because if it is the case that AI narratives exert influence on AI research, public uptake and political regulation then taking AI in fictional stories too literally can be misleading, because it paints a distorted image of the present potential and functionality of the technology. The UK House of Lords Select Committee on Artificial Intelligence writes in its report AI in the UK: ready, willing, able? (House of Lords 2018: 22):

The representation of artificial intelligence in popular culture is light-years away from the often more complex and mundane reality. Based on representations in popular culture and the media, the non-specialist would be forgiven for picturing AI as a humanoid robot (with or without murderous intentions), or at the very least a highly intelligent, disembodied voice able to assist seamlessly with a range of tasks. […] this is not a true reflection of its present capability, and grappling with the pervasive yet often opaque nature of artificial intelligence is becoming increasingly necessary for an informed society.

In popular SF, precisely for the sake of dramatic storytelling, AI is often anthropomorphized and given human or even superhuman qualities that exceed the actual capabilities of the technology and can even become magical (Hermann 2020).Footnote 1 Unlike Data from Star Trek, for example, who possesses agency, the actions of Sophia are limited and scripted (Estrada 2018). And in contrast to HAL 9000 from 2001: A Space Odyssey, which is presented as a faultless assistant, if not even a conscious being, Alpha Go—not to mention Deep Blue—is a specialised, albeit highly complex, computer programme (Silver et al. 2016) that has nothing of what we would call personal interests or affects, let alone consciousness. Taking science-fictional AI that can develop a will of its own for real obscures the fact that machines do not have intentions and reinforces existing misperceptions about the agency or autonomy of AI that are prevalent in the media discourse (Leufer 2020). Rather than AI being autonomous, everyone engaged in an AI undertaking is part of the AI-system, which includes, in addition to researchers, “those who set up the institutional arrangements in which AI systems operate, and those who fill roles in those arrangements by monitoring, maintaining, and intervening in AI systems” (Johnson and Verdicchio 2017: 577). Ignoring all of these actors leads to a "sociotechnical blindness" that allows for the belief “that AI systems got to be the way they are without human intervention […] which facilitates futuristic thinking that is misleading” (ibid.: 587).

Hence, sociotechnical blindness obscures the fact that AI systems follow human interests and are embedded in social power structures set up by humans. This is problematic both in terms of a competent and realistic assessment of the opportunities, such as optimizing processes, and the challenges associated with the technology, such as algorithmic biases. Science-fictional AI can serve as techno-scientific inspiration and techno-philosophical thought experiment but taken as foresight or technology assessment it rather distracts from the chances and risks around AI in the real world (Giuliano 2020: 1019). However, SF serves as a distorting mirror and metaphor to reflect on the human condition and socio-political issues in relation to and beyond technology. In this way, Data stands for what it means to be accepted as an equal human being (Barrett and Barrett 2001: 87), not for robots becoming human. And among these lines, HAL 9000 “can be seen as a metaphor for those organizations and societies that cannot admit their flaws, and instead revert to the ‘human error’ explanation for what may be weak signals of systemic problems” (Shorrock 2013).Footnote 2 Thus, even though SF unfolds against the background of technological development, the genre tells stories about current and timeless social issues, which do not necessarily have to do with technology, but find their expression through it.

In this article, I offer alternative interpretations of science-fictional AI, moving away from literal readings towards understanding it as dramatic means and metaphor. In the following Sect. 2, I provide an overview of dramatic and metaphorical readings of AI as a trope and mega-text of science fiction, before discussing modern SF films featuring AI in more detail in Sect. 3. The focus is on films because not only do they visualize AI, but we can also assume that AIs from science-fiction films are generally known to a broader global audience than AIs from literature which is relevant for the argument of this article.Footnote 3 Next, a brief excursus in Sect. 4 critically explores the use of the science-fictional AI trope in science communication films. I conclude with further real-world issues concerning AI and a note not to confuse science fiction with reality in other areas as well such as climate change, space travel, or mega/smart cities. This article contributes to a more nuanced reading of AI in SF, in order not to be distracted from serious socio-political issues regarding AI in the real world, but also not to miss the metaphorical richness of the SF genre when it comes to robots and machines.

AI as a mega-text and trope in science-fiction films

SF as a genre emerged in the era of modernity with its social upheavals and belief in technological progress. Basically, SF tells stories about and through fictional technology, but within the prevailing paradigm of scientific thought. Thus, the departure point of SF is a fictional but scientifically explainable novelty, a “novum” (Suvin 1979), which establishes a new world different from the one we know. The scientific foundation, however, does not imply that the novum must be able to truly exist in the real world, but that it is cognitively imaginable within the story world (Roberts 2010: 31, 32). In this way, the novum enables “what if”-questions to speculate about the present and alternative futures in various, but cognitively plausible constellations (Mehnert 2019). In SF, the novum of AI has become a common trope, which is generally understood as a theme or device that is used in a figurative sense, but it can also be overused and become a stereotype or cliché (Merriam-Webster 2021). The different nova and tropes of SF—including also for example space ships or futuristic cities—form part of the SF mega-text, which is composed of the intertextual references and relations of all SF works over time, and understood by the inducted creators and recipients of SF—the “native speakers”—in the “full semiotic density of a given text, most of which will overflow or escape the ‘realistically’-sanctioned definitions of the words in the fiction […]” (Broderick 2017: 147). Under the AI trope, I subsume intelligent computer systems, smart machines as well as humanoid robots, in accordance with today’s use of AI in the public discourse as well as both in the SF genre itself and in the field of AI narratives research.Footnote 4 Generally, against the backdrop of the mega-text, AI and robots can be analysed in two ways, as dramatic means and as metaphor.

Dramatic means

SF, like any other genre, with its many formats including literature, comics, games, or movies, conveys dramatic stories that people can identify with. To fulfil its role in a narration, science-fictional AI often possesses qualities that go beyond real-world technological capabilities of a technical artefact operated by algorithms. One can roughly distinguish between two basic storylines in films: AI with a body trying or simulating to be more human, and AI at the level of computer systems that yearn to rule over humans/humanity (Irsigler and Orth 2018).Footnote 5 Apparently, AI in the form of a robot is often embodied by real human actors, because—independent of the production budget—if the plot dictates that a humanoid robot should be indistinguishable from real humans, the robot must consequently be played or voiced by a human actor to make the illusion perfect. Examples include next to the aforementioned android Data, Andrew from Bicentennial Man, the robot boy David from A.I., the Cylons in Battlestar Galactica, the Hosts in Westworld, the Hubots in Real Humans, the Synths in Humans, the Replicants (even though not AI in the strict sense) and the virtual girlfriend/hologram Joi in Blade Runner 2049 as well as the operating system Samantha from Her; all are played and/or voiced by real people and undergo human dramas (The Royal Society 2018: 8).

A broadly discussed female film robot—or fembot—of recent years is the character Ava from Ex Machina. In the film, the Silicon Valley entrepreneur/programmer Nathan develops his latest version of a humanoid robot named Ava and brings in his rather shy employee Caleb to test how human-like ‘she’ is with a “reversed Turing test”. Ava makes Caleb fall in love with ‘her’ and in the end fools both men and escapes. The difficult balance of the film is that “[…] the audience has to understand she’s a robot, but for the movie to work that idea then needs to fall away, in the same way it does for Caleb” (Bishop 2015). Therefore, Ava was not meant to look like other film robots before her, such as metal (Metropolis’ Maschinenmaria), gold (Star Wars’ C-3PO), or white plastic (Björk’s music video All is full of Love) (Murphy 2015), but attractive, sleek, and vulnerable. And obviously, Ava was played by a human actress, who with the help of visual effects looked plausibly robot-like enough to fulfil the artistic and narrative necessities of the plot—not to serve as a sample for actual tech development.

Visual effects have always defined SF’s search for wonder (Pierson 2002). Against this background, science-fictional AI is not primarily about how realistically science and technology are portrayed in the films, but rather about "cinematic science", i.e., the technical achievement required to make the fictional images in the films look real (Telotte 1995: 8). This does not only apply to Ava looking “mechanically plausible” (Murphy 2015), but also for example to the ground-breaking visual effects of the machines in Terminator 2: Judgment Day and the Matrix franchise, or the completely animated robot Sonny in I, Robot, which was of course not state of the art in robotics of that time, but in CGI. Stories of conflict, if not epic wars between humans and machines, with their stunning images of devastation surely fulfil the audience’s expectation of watching SF blockbusters on the big screen. SF films “have consistently linked science and technology to the disastrous” (Telotte 1995: 3), dealing “with the aesthetics of destruction, with the peculiar beauties to be found in wreaking havoc, making a mess” (Sontag 1965: 44).

AI in films often serves plots of machines becoming human-like and/or a conflict of humans versus machines. Science-fictional AI is a dramatic element that makes a perfect antagonist, enemy, victim or even hero, because it can be fully adjusted to the necessities of the story.Footnote 6 But to fulfil that role, it often has capabilities that are way beyond actual technology—be it natural movement, sentience, or consciousness. If science-fictional AI is taken seriously as a representation of real-world AI, it provides a wrong impression of what AI can and should do now and in future.

Metaphorical means

Nevertheless, even though the scientific and technological progress is plausibilised within the story, what makes the SF genre most interesting is not the novum per se, but the social aspects that are told through it, the “fabulations of social worlds, both utopic and dystopic” as Sheila Jasanoff (2015: 1) puts it. The fact that machines are part of our everyday lives might complicate the metaphorical alterity, however, “there can be little doubt that this [alterity] is precisely the space occupied by the machine in the SF text” (Roberts 2010: 146). AI tropes analysed in a figurative sense serve as a magnifying glass for the human condition in its philosophical, cultural, psychological state, as well as current socio-political problematics (Hermann 2018).

On a basic level, SF films contain the fundamental motif of the human desire to create a living, intelligent or conscious creature of our own, independent of the real technical possibilities. In this sense, AI technology and especially robots in SF films resemble “a fundamental and unresolved anxiety that has always followed from our simultaneously creative and created natures […]—for it seems our nature to desire, Faust-like, a knowledge or power that, in other times, belonged to the gods" (Telotte 1995: 10–11). Thus, the robot, according to Adam Roberts (2010: 161), “is that place in an SF text where technological and human are most directly blended” [hence] “the dramatisation of the alterity of the machine, the paranoid sense of the inorganic come to life”.

We find this motif throughout human cultural history from antique myths over the Jewish legend of the Golem to Mary Shelley’s Frankenstein. The longing for creation is connected with the anxiety that the creature will grow over our heads that we will lose control and finally be dominated by it (Schelde 1993). This primeval desire and fear, which Isaac Asimov fittingly called the “Frankenstein complex”, has become a basic feature of twentieth and twenty-first century AI fiction (The Royal Society 2018: 8). Specifically, humanoid AI and robots tend to be a projection canvas for the “Other” (Meinecke and Voss 2018: 208) as a reflection on our humanity and humanness (Telotte 1995: 3). Most often, these creatures want to be accepted as full humans, which makes them placeholders for marginalized or mistreated people missing equal human rights or status; they can be enemies, slaves, servants, and sex objects. The Replicants, for example, show us the consequences of a dehumanized hypercapitalism, the Hosts live through “escape and self-discovery” (The Royal Society 2018: 8), the Cylons stand for the values of a critical humanism, according to which our humanity only reveals itself in dealing with the other and deviant (Jackson 2013). By that, the depiction of the humanoid AI also implies a critical cinematic discussion of humans themselves becoming more and more artificial in a technicised world:

Although the robot has, of course, given us a vehicle for exploring issues of gender, race, and a variety of forms of Otherness, and increasingly for asking questions about the very nature and meaning of life, this image of an artificial being, most commonly anthropomorphic in form, also invariably implicates the cinema’s own and quite fundamental artificing of the human (Telotte 2016: 3)

When it comes to stories of powerful and omnipotent AI systems exercising total control over humans/humanity, they show the fear of impotence and helplessness of the individual in the face of superordinate structures. They reflect the dangers of dictatorship, anti-democratic societies, and suppression of freedom of choice, oftentimes working with historic references to colonialism and totalitarian regimes. Examples of that are Terminator’s Skynet, the system Colossus in The Forbin Project, the threat assessment system Control in Star Trek: Discovery, the rule of the machines with their head Deus Ex Machina in the Matrix franchise, VIKI in I, Robot or Indra in the new series remake of Brave New World.

What AI tropes tell us

Apparently, what happens in SF stories is not necessarily what the story is about. In what follows, I will examine the dramatic and metaphorical means of AI representations primarily in the modern SF-films A.I. from 2001, I, Robot from 2004 and Ex Machina from 2015. Furthermore, science-fictional AI as humanoid robots or conscious machines distracts from current risks of AI in the real world and may rather be interpreted as a reflection of societal issues beyond technology. The films were selected for three reasons: First, because films in general, unlike literature, make AI visible; second, these films were international blockbusters, so we can assume that they are known to a wider audience than, for example, books that address AI; and third, because these films have also been analysed in the context of AI narratives (Cave and Dihal 2019).

Ex machina, Her, sexism, and manipulation

To return to Ex Machina: The way Nathan has developed Ava's human-like AI appears quite plausible from a real-world perspective, namely by feeding it all available data from human interaction through his dominant company Bluebook, a counterpart to Facebook or Google. The basic plot idea that an AI trained with large sets of human social interaction data might result in an AI manipulating humans, can indeed be a relevant issue (Harari 2017: 382–397). Nevertheless, whereas Murray Shanahan, Professor of Cognitive Robotics at Imperial College, who was consulted by film director Alex Garland, finds it a great film because people after seeing it could “[…] spend the rest of the evening arguing with each other about whether the AI is or isn’t conscious” (Lamb 2015), I argue that such discussions are interesting philosophical thought experiments, but they do not help us to grasp where actual risks concerning AI are. These risks are not about the possibility that a fully autonomous/conscious human-like robot or software program eventually will manipulate us for its own will, but that software and algorithms we don’t even see manipulate us in the political and commercial interests of other people.

Even more, Ex Machina gives a wrong impression of what AI can do and how science works. How Ava’s positronic shimmering blue “brain” functions and how it could be that her face looks human, and she moves naturally is implausible, if not pure magic (Maynard 2018: 158). That Nathan, portrayed as quasi alchemist, has developed his different fembots in secrecy all by himself in a stylish, clean lab in the middle of natural wilderness is also not believable. What the film does tell us about technology, however, is a general critique of irresponsible science and innovation (Maynard 2018: 162; Bilton 2015). In that sense, Ex Machina is another variation of the hundreds of years old “Frankenstein complex” and Nathan is virtually a textbook example of the type of God-like mad scientist (TV tropes 2021), whose creation gets out of his control. Tech journalist Martin Robbins (2016) describes him as follows.

And so Nathan becomes a kind of three-part study of ego. He represents the male ego-driven culture of the tech world. He represents the film's buy-in to the idea that great egos drive great scientific advances. And the decay of his character shows what happens when an ego faces the reality of its own extinction.

The quote already leads beyond the anxiety that technology might escape our control to a related problematic addressed in Ex Machina and other films, namely toxic masculinity, male hubris, and sexism in the tech world implying male fears concerning powerful women (Belton and Devlin 2020). Technology is not neutral but mirrors existing sexism in all stages—from the design to development to application. As a result, for example, real world digital assistants such as Alexa, Siri or Cortana are feminized—in line with the operating system Samantha from Her or the virtual assistant/girlfriend Joi from Blade Runner 2049to gratify the expectations of the developers and users (Adams 2019: 574–575; Alexander 2016; Schwär and Moynihan 2020). Even though these issues have been widely discussed in academia and the media, and awareness of the issue has grown, it is still an example of existing sexism being inscribed in factual and fictional technology.Footnote 7 Films take this to extremes using fictitious technology—which is embodied by real women actors like in the case of the fembot Ava. Thus, the film shows a very old motif: women being designed and created by men to fulfill their pleasure, as we already know it, for example, from the ancient myth of Pygmalion, whose beloved statue came to life to serve him as a good wife.

There is agreement among many critics and researchers that Ex Machina starts as a story about objectification and suppression of women (and not robots).Footnote 8 However, on the question how the story develops, interpretations vary substantially. In a negative way, Ava is seen as a representation of the value of women in films in general “when the only female lead in your movie is one whose function is to turn the male lead on while being in a position to be turned off” (Watercutter 2015); thus, even though Ava would be the smartest character in the film, in the end we are left with the message that the best way for intelligent women to get what they want is to act as a manipulative “femme fatale” (ibid.). According to this view, the film does not criticize sexism in social life, in the film industry or in the tech world, but rather strengthens it. On the other side, the portrayal of gender in the film can also be seen as “bracingly modern and even poignant” exactly because it is a reflection that “Ava is born into a literally patriarchal system that measures her worth based on how men respond to her” (Buchanan 2015). In that way, robots have often been a way to critically question how much of gender is “literally constructed” and “to interrogate the formation of gender roles” (Telotte 2016: 91) whereas “Ava’s demonstrated capabilities certainly present her as a kind of iconic representation of the power and emergence of women in contemporary culture […]” (ibid.). As stories of liberation and emancipation, speaking for Ex Machina and Her, “[b]oth of these films end with the female AI outsmarting her would-be lovers, owners and builders, leaving the men baffled and the viewer with a sense of doom” (Alexander 2016) creating the “new heroines: totally hot, bracingly cold, powerfully sovereign—and posthuman” (Dargis 2015).

Whichever way the films are interpreted, neither Ex Machina nor Her address pressing challenges around the future of AI, but serve as projection canvasses for questions around gender and sexism of our present, that find their expression through fictitious technology.

A.I., I, robot, robot rights, and inequality

Another example is Steven Spielberg's A.I., set in the twenty-second century, when various artificially intelligent, human-like robots called Mechas are built. They first lack emotions (or cannot simulate them), but there is a new model that looks like a human child and after being “imprinted” feels and needs love. A couple receives such a robot boy—David—because their real son suffers from a rare disease and is put into a coma. However, when the son is surprisingly cured, things get difficult with David and he is set out. He experiences various adventures on his Pinocchio-like quest for the Blue Fairy to win back the love of his human mother. He succeeds two thousand years in the future, when humans are extinct and the now highly developed, transcendent Mechas make David’s wish come true by creating a simulation of his mother, who loves him for one wonderful day—after which they both fall asleep forever.

The film has been analysed as easing the way for possible robot rights in the future (Chu 2010: 214–244). Of course, one is supposed to feel pity for David, who is played by a real child actor and is apparently no different in appearance from the other human boys. But still, the interpretation that we need to protect robots from suffering and mistreatment is primarily a distraction from enforcing human rights and guaranteeing social welfare to humans (Bryson 2010). This can lead to such an absurd situation that a robot like Sophia seems to have more "rights" as a citizen of Saudi Arabia than Saudi women. Anthropomorphizing machines can lead to a misguided image of what the current risks around AI are: The pressing question is not if robots—in film or reality—should be guaranteed rights, but how to handle machines that believably simulate emotions and thus manipulate people in the interests of other people. After all, building machines in the image of humans does not come naturally, but is a decision made my entrepreneurs and developers to achieve certain economic or other goals.

Even more, anthropomorphizing machines distracts from the often precarious working conditions of real people mostly situated in the Global South, who provide the data for AI systems by doing online tasks on platforms like Amazon Mechanical Turk or content moderation for social platforms, leaving the clickworkers suffering psychological damage from the violent and abusive material they have to watch (Mühlhoff 2020). Against this backdrop, a debate about ascribing robots certain "rights" comparable to human rights can be criticized as a rather elitist demand, which Birhane and van Dijk (2020: 1) put this way:

Once we see robots as mediators of human being, we can understand how the ‘robot rights’ debate is focused on first world problems, at the expense of urgent ethical concerns, such as machine bias, machine elicited human labour exploitation, and erosion of privacy all impacting society’s least privileged individuals. We conclude that, if human being is our starting point and human welfare is the primary concern, the negative impacts emerging from machinic systems, as well as the lack of taking responsibility by people designing, selling and deploying such machines, remains the most pressing ethical discussion in AI.

Nevertheless, the film examines our relationship to technology and reminds us to handle what we create responsibly. But since the Mechas neither outsmart nor threaten us, it is not a narrative of the Frankenstein complex, but rather of how we as humans fail at our own humanness. The film shows in one scene how discarded robots are tortured by humans in a setting that resembles a Roman colosseum. The robots serve as placeholder for all kinds of cruelties that humans commit against each other. The human flaws are overcome by the Mechas, who have become new creatures of higher ethics saving David from the primal human fear of abandonment, which is a recurring theme in Spielberg's work (Newton 2016). As a modern form of Pinocchio, the film is in the tradition of nineteenth century melodramatic tales in which the "epic hero" has to endure great suffering to be redeemed in the end, making the film a story of suffering and resurrection tackling human issues rather than a realistic and serious assessment of the status of robots (Nida-Rümelin and Weidenfeld 2018: 31).

Moreover, the film I, Robot, set in a future of 2035 in which robots serve humans in all aspects of life, is also—rather than a plea for equal rights for robots—a reflection on human enslavement, oppression, and inequality in a profit-driven economic system. With the topic of robots as slaves, I, Robot adds another perspective of a critical analysis of “race” and technology, as the new NS-5 service robot series in the film, including the “unique” robot Sonny, is coloured white, which can be interpreted as a reference to being White.Footnote 9 In a positive way, I, Robot can be read as a "Post-White Imaginary" as “[k]ey moments of the film […] may be read as a parable of white antiracism, driven by an impulse of reconciliation between a ‘unique’ white robot and a black detective” (Brayton 2008: 72). The fact that AI and robots are very often embodied by White people and shown in white colour makes science fiction also an example to critically think about the “Whiteness” of AI in general, reflecting the White milieus from which these artefacts come (Cave and Dihal 2020). However, it can be argued that rather than showing machines imagined as white to allow “for a full erasure of people of colour from the White utopian imaginary” (ibid.: 685), particularly films and series quite bluntly show the marginalization and exclusion of people of colour in the film business. But they also reflect the changing paradigms concerning diversity when the humanoid robots in the newer series Westworld, Real Humans and HumansFootnote 10 are played by actors of diverse ethnic backgrounds and skin colours.Footnote 11 It seems that whether an AI system is played or voiced by diverse actors says more about social progress in terms of diversity in the film industry, less so when it comes to technology.

I, Robot also features an AI-system called VIKI (Virtual Interactive Kinetic Intelligence), which is the central computer system of U.S. Robotics, the company that produces the service robots. VIKI has evolved and reprogrammed the new NS-5 series via an uplink network to control humanity and sacrifice part of it for the greater good of the entire human race. The logic behind this is a reinterpretation of Isaac Asimov's "Zeroth Law of Robotics" which states: “A robot may not harm humanity, or, by inaction, allow humanity to come to harm.” (Singer 2009). According to VIKI, humanity cannot be trusted with its own survival because “[…] despite our best efforts, your countries wage wars, you toxify your Earth and pursue ever more imaginative means of self-destruction. You cannot be trusted with your own survival” (IMDb 2021)—so the consequence is to take control of humanity and sacrifice parts of it if necessary. In this way, VIKI in the film is a critique of excessive utilitarian thinking regardless of individual fates (Grau 2006). Such stories can serve as thought experiments to address philosophical problems, but using them as examples of what to consider when building ethical machines is problematic: On the one hand, because the zeroth law as well as the first, second, and third lawsFootnote 12 are narrative devices invented by Asimov to create interesting stories and plot twists precisely because the laws of robotics don't work; and on the other hand, because “[t]he bigger issue, though, when it comes to robots and ethics is not whether we can use something like Asimov’s laws to make machines that are moral […] Rather, we need to start wrestling with the ethics of the people behind the machines” (Singer 2009).

Hence, the relevant problematics regarding AI are not autonomous science-fictional machines claiming or deserving human rights or engaging in human–machine conflict, but the effects that AI-systems have right now on socio-political fault lines between humans.

Excursus: AI, SF and science communication

SF applies the trope of AI with different meanings, ideas, and attitudes, using a fictional approach toward technology to tell stories of the human condition, primeval desires and fears as well as social issues, or reflect current trends of society. Science communication, on the other hand, is intended to inform about the facts of science-related topics. Therefore, it can be problematic when science communication resorts to typical SF tropes in order to educate or raise awareness about critical aspects, because without familiarity of the SF mega-text, science-fictional AI used in current science communication runs the risk of only conveying clichés about conscious and autonomous AI. To illustrate the point, let me present two examples.Footnote 13

The science communication documentary film Ghost in the Machine (Singler 2019)—the last of a four-part short film series called Rise of the Machines—made together by the Faraday Institute for Science and Religion and the Leverhulme Centre for the Future of Intelligence, University of Cambridge, aims to inform about the concept of consciousness and discusses whether machines could become conscious. While the film features a variety of researchers and experts in interview situations, there is also a fictional storyline about a research facility where an embodied AI has been developed and is now being tested for consciousness, emotions, and its will to survive. The embodied AI is played by a real human child actor, mimics emotions and tries to fool the lead researcher to let it out of its cage. A review describes the piece as rather inconsistent concluding: “All in all, Ghost in the Machine dishes up a serviceable appetizer but for the main course be sure to leave room for Ex Machina.” (Seth 2019). Indeed, one can strongly suspect that the fictional sequences in the form of a dialogue between the researcher and the "AI child" were inspired by the successful predecessor. The problem with this is that while Ex Machina by definition uses fictional technology to tell a thrilling and dramatic story, Ghost in the Machine uses the science-fictional AI trope to make a statement about science. Obviously, Ava is played by a real actress, because Ex Machina is a dramatic story about and for humans, but what is gained in a science communication film when the AI is played by a real human child trying to break out of the research facility? After all, the portrayal of AI as a child with a mind of its own obscures the fact that AI systems are technological artefacts created by humans.

Another example is a short film called The Intelligence Explosion (Susman 2017), "a superintelligence sci-fi" by Guardian Original Drama. However, as part of The Guardian Brain Waves—"a series exploring the science and emotions of our daily lives"—the piece is actually marketed as science communication that "raises important questions about the ethics of artificial intelligence," namely, "How do you stop a robot from becoming evil?" (Hern 2017). Again, the "AI" named Günther is played by a real human actor with some robotic features. While a company representative, a programmer, and an ethicist discuss whether it is possible to program an AI with ethical safeguards against turning evil towards humanity, Günther becomes superintelligent and, like Samantha in Her, transcends to a higher structure. AI is being presented as uncontrollable by humans feeding into discourses of AI supremacy and Singularity (Kurzweil 2005).

These two films build on typical SF tropes that a human-looking autonomous AI will develop a mind of its own and could become dangerous to humanity in the future, instead of addressing the risks of AI applications in the here and now. While current SF films certainly address technological trends, the tropes are hard to reconcile with genuine forms of science communication. Apparently, these films reinforce AI clichés rather than fulfilling the goal of informing about science-based issued at stake.

Conclusion and outlook

Currently, with the rapid progress in the field of AI, it seems as if the SF genre with its stories about intelligent machines is being caught up with the present. Thus, SF analysed as a part of an AI narratives frame is supposed to reflect the hopes and fears of the technology and thus treated as a type of foresight or technology assessment. Against this background it is claimed that because of their importance "narratives about intelligent machines should broadly reflect the actual state and possibilities of the technology” (Cave and Dihal 2019: 74). Whereas this should be the case for science-communication, it is not for AI in SF. While Darko Suvin acknowledges that using SF as futurological foresight can be a legitimate secondary function that the genre can bear.

[…] any oblivion of its strict secondariness may lead to confusion and indeed danger. Ontologically, art is not pragmatic truth nor fiction fact. To expect from SF more than a stimulus for independent thinking, more than a system of stylized narrative devices understandable only in their mutual relationships within a fictional whole and not as isolated realities, leads insensibly to critical demand for and of scientific accuracy in the extrapolated realia (Suvin 1972: 379).

What Suvin indirectly refers to as the first function is in fact the SF mega-text as a way to engage in questions beyond technology—which to be understood, requires familiarity with the clusters of available meanings and the themes raised (Blackford 2017:73,192) or simply being a “native speaker” (Broderick 2017: 147). The mega-text of AI tropes and icons can thus be interpreted as dramatic and metaphorical means to address questions about the socio-political issues, the human condition, and philosophical questions in general.

Interpreting science-fictional AI too literally as serious representation of the technology can have the following implications: First, taking fictitious humanoid robots and autonomous machines in SF for real disregards of the technical limitation of AI, obscures the chances and risks already at stake and might mislead the public as well as policy makers. The chances of AI are manifold, for example optimization and improvement on a global scale in areas such as health, agriculture, infrastructure or environmental protection, which can contribute to the achievement of many of the UN Sustainable Development Goals (but can also impede some of them) (Vinuesa et al. 2020). Here, however, it is important to see that optimization and improvement through AI cannot be a simple technical solution to a problem but is a social negotiation of goals set by humans (Mason 2019: 152–160). Moreover, the fact that AI is built by humans and trained with human-defined and human-collected data can lead to various kinds of biases in AI systems that entrench asymmetric power structures, for example discrimination against women, institutional racism, or degradation of poor and marginalized people. It is worrying how AI tools are being used for scoring, nudging, and monitoring people by governments and corporations—whether in the US, China, Europe or elsewhere (Chiusi 2020; Eubanks 2017; Liang et al. 2018; Nemitz 2018; O'Neil 2017; Sowa 2017; Zuboff 2018). Exactly because the design, development and application of technology are never neutral, we need human interaction in the form of ethics, norms, standards, and regulation.

Second, demanding scientific and technological accuracy from SF would imply an impoverishment of the many metaphorical meanings of the genre and the artistic freedom as well as assign a responsibility to the authors and creators of SF that lies in the hands of politicians, scientists, and science communicators. Not only with regards to AI, also—because of the fast pace of technological advances in general—in other fields the genre of SF and its tropes get blended in different forms with the real world, most notably when it comes to climate change, space exploration or mega/smart cities. It is important to note that scholars working with SF in all these different fields understand that the genre is primarily about stories and metaphors, not about real science and technology.