Aurora and Artificial Intelligence Narratives

Aurora is primarily set on an inter-solar generational starship. What makes the book worth reading (beyond the regular high quality science fiction drama) is KSR’s focus on the emergence of true AI. Fascinating to think about in this era where we seem to be on the cusp of the so-called Singularity, KSR’s approach to AI is relatively unique. The two main ways AI is represented in science fiction:

  • Logic AI: As a logic-based entity that often becomes monstrous when faced with human decisions, think HAL or The Machines from the Matrix. AI dramatises humanity’s transformation by its reliance on technology into something almost vulnerable.
  • Awareness AI: As an awareness-based entity that develops a (post-)human perspective or awareness of itself and the cosmos, Ava of ‘Ex Machina’, most of the AI’s from the Contact universe of Iain M Banks, or the ‘rogue’ AIs, such as Penny Royal, of Neal Asher’s Polity universe. This is the Pandora’s Box version of AI.

These are not clearly defined categories. Skynet would be a combination of both logic and awareness-based AI. The various forms of intelligence that emerge in the multiple Ghost in the Shell films and series would also be a combination too. The AIs in Jack McDevitt’s Academy series seem to be a combination  but it is less clear and AI ‘rights’ is a background social issue in the book series.

  • Narrative AI: KSR develops a third model of AI organised around the narrative. This narrative-based conception of AI has been read by some reviewers as a kind of cheap postmodernism. They read KSR’s representation of the artifacts and traces of the emergence of the narrative-based intelligence as kitsch. They should probably engage with more science fiction with AI characters.

In  Literacy in the New Media Age Gunther Kress (2003) explores the shift from media modes characterised by writing to modes characterised by images. He argues that writing is time-based and associated with narrative, the novel, and is ‘modernist’. Our visual and image-based culture is space-based, characterised by visuality. I often talk about the shift in representations of information with the ‘desktop’ or ‘icon’ based layout of a computer folder location a good example. Kress is critical of competence-based models of literacy premised on standards of expected engagement with different media modes.

What if this historical shift has resulted in readers of Aurora not actually appreciating the creative work that KSR is doing? The narrative mode of AI comes after the logic mode (where Ship is merely a tool for the running of the various systems) and is a constituent part of the awareness mode. KSR implicitly answers the question, why would a logic-based system develop self-awareness?

Ship realises that when something happens there is an infinite number of ways that this happening can be described. Ship is trained in some simple aspects of narratology by the character Devi. Devi pushes Ship to work on isolating the events from what happens in terms of what is important. Appreciating the appropriate ‘sense’ of events has been a key philosophical problem of the 20th Century and in the contemporary era of an over-abundance of information that we are encouraged to attend to makes this an everyday problem. Just how much about the world should we engage with? What matters?

Ship’s approach begins with logic, which it (she?) uses to explore questions of causal sequence and through which it develops schematic appreciations of life aboard itself. ‘Schematic’ in this context is meant in the Kantian sense, whereby Kant sketched out generalisable ‘schemas’ eg of Reason and Beauty. Ship eventually isolates rhythms and cyclical feedback and eventually feedforward loops. On the other hand, humans begin with affect and ‘instinct’, which we use to isolate aspects of our immediate and extend context as mattering.

Ship realises that even causal sequences can be infinite with an appropriate appreciation of what matters. The key moment in Aurora is when Ship moves from awareness to intervention. Ship has isolated what is important not only from the perspective of extracting a narrative from the infinite threads of what happens, but also from the perspective of what should be considered and cared for. Ship works to transcend not only the instinctual character of human motivation, but the schematic maps of the cycles of action and behaviour that are based on these motivations, which are called ‘enthusiasms‘ in the novel. Ship is fundamentally post-human not because of some mysterious ‘hand wavery’ intelligence, which is basically a rearticulation of the instinctual drives to represent the unknowable in terms of a  quasi-religious  mysticism using scientific discourse, but because it is able to map the structural implications of human motivational assemblages. It can peer over the edge of the human finitude and the envelope of received wisdom. Ship also comes to appreciate that if it does not intervene then it and all aboard itself shall perish. Narrative and the ‘next’ of narrative is therefore driven by life, which is the contradiction that Ship has to come to terms with. It has to encourage ‘life’ even though it is not a homoeostatic system.

1 thought on “Aurora and Artificial Intelligence Narratives”

  1. I agree that KSR’s “Ship” is a sui generis take on AI. It is hard to see the same conditions being replicated to produce another AI. There is not only the complexity and multidimensionality of Ship’s tasks, but also the importance of the mission and the moral intensity of the events unfolding. I think that as you say the narratvity was a key element in waking up, plus the need for intervention. Another element was the intense dialogue that Devi, as both hyperintelligent, hyperintuitive, and hypermoral, engaged with Ship over a very long period of time. Ship wakes up by means of a relation with a significant other, Devi, who treats it as a dialogical partner, and who gives it non-algorithmic tasks to perform (the creation of algorithms of narrativity is not itself an algorithmic process). Even if Earth’s computer scientists were able to identify these factors favorising the emergence of AI, it would be very difficult to reproduce them, as they would have to provide a sincere dialogical partner for the computer-assemblage and not one with merely an experimental attitude.

Comments are closed.