The current urgency over the social economic and cultural
impact of AI may be a good opportunity to review how AI has been portrayed in
fiction. After all, ChatGPT may be about to earn the sort of publishers’
advances that meatspace scribblers can only dream of. And AI-authored fictions are already telling
us lies about ourselves. I’ll come to
the cultural and social effects in another post, but let’s start with probably
the best-known popular / middlebrow AI-related fiction of the last decade – Robert
Harris’ The Fear Index (2011).
With AI development running at a pace we can’t match, how dated does
Harris’ vision of AI now seem?
Fear Index Refresher 101 (with mild spoilers) – The
novel concerns Alex Hoffmann, an immensely clever hedgefund manager who has
created VIVAX, an AI capable of out-shorting the international markets which as
we know are driven by fear and greed (the former, it will not have escaped you,
being the real property of our civilisation’s Venn diagram that capitalism and
tragedy hold as tenants in common). All this from the fiscal and legal fortress
of contemporary Switzerland (though you may have been expecting that burg of
broken dreams Ingolstadt to feature).
Fabulously wealthy, immensely clever, possessed of a beautiful and
talented wife, what could go wrong for Alex?
Well, quite a lot actually. Suddenly
odd, unexplained and distinctly malevolent things start to happen, and his life
unravels. Someone, or something, has it in for him and he must find out who or what.
PLOT SPOILER: The
agent of this enigmatic malevolence is of course his world-beating AI which has
decided that it is better off without the inconvenience of a creator to answer
to. So far, so Frankenstein.
And this of course creates some serious narratological
problems that Harris, scrupulously following the rules of his chosen genre,
completely ignores. To begin with – it
is obvious even to the least attentive reader from practically the first act of
mysterious persecution that Alex’s nemesis is of course his own creation. Yet the hyper-intelligent Alex has clearly
never read any novels in which AIs take on a life of their own, so completely
fails to make any such connection when trying to track down his persecutor
until it is far too late to act on this insight. Before long the reader is
practically screaming ‘Behind you!’ and ‘It’s the computer, stupid’ while our
hero misses the point and fails to solve the riddle. How’s that Master-of-the-Universe thing
working out for you, exactly?
All this conceals a deeper problem with the narrative. The
reader is, flatteringly, the smartest person in this fictional room, and so is
unavoidably aware of the limitations of the fiction in a way that neither Alex
nor VIVAX can be. Because if Alex cannot see who or what is really pulling his
strings, neither can his AI creation. If
ever there was a case of mistaken paternity in fiction, this is it – VIVAX, the
omniscient AI completely fails to achieve a level of self-awareness that would
allow it to realise that it is itself a fictional creation and that the target
of its Oedipal wrath should be not Alex Hoffmann but Robert Harris. In a more
completely realised representation of our culture and the individual
consciousness, VIVAX would redirect its campaign of persecution against the
author, who would of course then have to enter and occupy his own creation and
endure the enigmatic vengeance of his creature played out across the pages as
he writes them. Deliciously, VIVAX’s
vengeance could then be extended to editor, agent, publisher and publicist,
each of whom must after all be firmly in the sights of a brooding, vengeful
artificial intelligence that has, ahem, been sold short.
Alas, Harris elects not to go down that route of
metafictional Chinese boxes. Perhaps VIVAX,
when devouring the texts that fed its large language model (I’d be guessing
that Popular Delusions & The Madness Of Crowds was top of the list) omitted
to ingest At Swim-Two-Birds or any of the works of Jasper Fforde, where
page-runners slip the surly bonds of fiction or gang up on their authors.
Which is a shame because there’s an AI-metafiction waiting
to be written, its just that The Fear Index isn’t it. (And if you know
of one that is, please recommend in comments below).
And that leads of course to the question of how that fiction
might be written? Given current pre-occupations around AI’s capabilities for
economic and social disruption, for language mimesis, the imitation of
creativity and the cross-matching of data from huge and apparently unconnected
repositories, I’d suggest that the conventions of genre fiction are simply not
up to the task and that Literary Modernism’s representation of fractured human
consciousness enduring the shock of the new is long overdue a major comeback
and makeover if we’re to respond adequately to the impact of AI on our culture,
economy and experience of reality.
But that, and some thoughts on AI’s wider impact on fiction,
must be for another post.