David McFarlane joins the #SCiFi, #AI and the Future of War series with this discussion of the movie Chappie, directed by Neill Blomkamp. David argues the movie usefully prompts us to think through how our perceptions of AI will shape relationships, and challenges humans to consider that the notion of unintended consequences in dealing with AI may be different to previous developments.

There are a number of sci-fi movies that present a future where the AI we have invented now rule over us and are competing with the human race for scarce resources. Chappie, on the other hand, envisages a future where the first invention of advanced AI takes the shape of a single predominately humanised robot. Chappie is the first robot to exhibit consciousness in a world where robots have taken over the majority of police work in Johannesburg. This happens after he is reprogrammed by his creator Deon to ‘think and feel.’ Throughout the film, Chappie grows from learning to speak and paint to an adolescent thug, to a fully sentient adult aware of his mortality. Deon’s colleague Vincent sees thinking robots as unnatural and a risk, and so between the two, we can explore what may potentially be the two schools of thoughts on advanced AI.

Deon (and by extension Chappie) is initially kidnapped by three criminals, Ninja, Yolandi and Yankie, who aim to shut down the city-wide robotic police force. They then adjust their plans with the hope that Chappie will catalyse their meteoric rise to criminal stardom. In early interactions between Deon and Chappie, Deon makes Chappie promise that he will not participate in any illegal activity. Ninja, or Chappie’s ‘Dad,’ learns to bypass this promise by reframing his requests. He convinces Chappie to steal a car by claiming the car was his to begin with, and somebody else stole ‘Daddy’s car.’

Consequently, Ninja convinces Chappie that stabbing someone is not a crime; and instead was a beneficial activity in that it would put them to sleep and feel nice. Chappie refuses to participate in a heist when asked, in line with his promises to Deon. This is until he becomes fully sentient and realises that he will die when his battery runs out. In response to this, he utilises the Internet, and every bit of information humanity has known to formulate a way to transfer consciousness. Despite his promise, this endeavour necessitates and ensures his participation in the heist.

A stark opponent to everything Chappie represents, Vincent shuts down the robotic police force to eliminate him. The movie culminates in the ‘Moose’ fighting chappie. The Moose is an AI weapons system that is still operated by a human (Vincent), comparable perhaps to drone pilots today mixed with advanced real-time display. This likely reflects where AI will take us in the near future in a military context. The movie concludes with Deon being shot and Chappie hoping to save him. Chappie engages in a revenge mission, acting on emotion and disregarding any care for the rule/processes of law or sentencing towards Vincent. He does not, however, execute him. He then begins transferring Deon’s consciousness into another AI shell/body.

Discussion: The film opens with two telling sentiments, the first “It’s too early to tell how this will all play out. I didn’t think this would happen in my lifetime, but it is happening.” The second is “when we look at evolution, it’s not surprising that Chappie has taken this turn.” These two thoughts are poignant considerations in any discussion of the future of AI, and in looking at the developments in Chappie that we currently perceive to be impossible.

The beginning segments of Chappie explore an issue that should always be a consideration in employing AI technologies: they can be stolen and/or compromised by potential adversaries. Specific to this dilemma is how susceptible organisations are to internal compromise, as seen through Vincent decommissioning the mechanised police force. The very technologies we develop to enhance our lives or capability could be the ones that are weaponised or compromised to defeat us. Now, it could be said that this is inherent in any new technology we develop. However what separates AI from the past is that when an asymmetric weapon is capable of consciousness and organic thinking, of problem-solving faster and evolving faster than us, how do we stop it? The answer that springs to mind is more or different AI, which is again vulnerable to infiltration as discussed earlier.

When Deon proposes the development of sentient AI to his superior, he is met with this response: “You just came to the CEO of a publicly traded weapons company to pitch a robot that can write poetry. Why do we want it?” This retort resonated with me throughout the film and prompted a few core questions. Assuming of course that there is a future where sentient AI technology is possible, we should not ask ourselves if we should pursue such a reality, but rather why we should endeavour in such a pursuit? It is prudent that we take an effects-based approach in any such development, and always ensure the positives outweigh both potential and unknown risk. Furthermore, if we were to develop AI capable of sentient thought, should they then be entitled to the same rights we generally bestow to other intelligent beings? Finally, in the case of looking at a brilliant lone-actor in Deon, will we be powerless to stop and/or monitor said development?

Also explored in Chappie is our early and sustained empathy for him. This is especially prominent during his ‘infancy’, where we see him abused, neglected and manipulated by those closest to him. His first word is “watch”, a word he learned solely through the imitation of his creator. These developments not only reaffirm the risk present in AI being dangerously misused or mislead but also raise another concern. For many humans, empathy, trust, emotional investment and indeed emotional attachment are integral parts of our lives. These facets of human nature combined with forms of AI that are increasingly integrated into our lives could reduce our ability to be impartial and critique. What if the benefits of AI become threat or risk at a slow enough rate that that attachment and trust blind us? Chappie’s increasing intelligence throughout the movie is a telling narrative. Research has indicated that that is where the future of AI lies. AI, as Chappie did, will most likely be able to evolve itself through a process called recursive self-improvement; their ability to continually make their software better.

As with AI in general, the ideals within and explored around chappie must originate from a programmer/programmers. This dictates that it is the programmer’s agenda that decides what consciousness could look like. This includes but is not limited to the rights and wrongs associated with such consciousness, what is desirable in values, interpretations of laws and so on. It goes without saying that this reflects a risk in both the military and civilian application of AI.

Pilot Officer David McFarlane is an officer in the Royal Australian Air Force. The views expressed are his alone and do not reflect the opinion of the Royal Australian Air Force, the Department of Defence, or the Australian Government.