There has been much talk in recent months of the evolution of artificial intelligence. Even the oft-derided idea of a machine uprising has been discussed, and by serious scientists no less! These discussions got me thinking about the idea of AI and creative writing. What if computers could write, not just simple, pre-programmed sentences, but entire novels, stories, plays, and screenplays?
The French director Bertrand Tavernier (after the work of sci-fi novelist D. G. Compton) imagined such a world in his 1980 film Death Watch – a world where people no longer die of natural causes and where literature is written by machines. There is much critical (and other) literature on the subject of death in the film and its source text. Instead of dealing with death in Death Watch, I focus on the creative, generative aspect – writing, the idea of automated fiction.
Katherine Mortenhoe, the film’s female protagonist, is a ‘programmer’ of romance novels. She is a writer – her sentences, similes, metaphors, and dialogue, however, are programmed. She writes code. Her computer, affectionately known as Harriet, then produces novels that are guaranteed to sell. On one level, the notion of a fiction ‘programmer’ is a clear acerbic comment on the formulaic nature of narrative-building in the genre others have called ‘chick lit’, or a wider comment on the nature of mass-market literature in general.
But more than this, the exploration of the idea of automated fiction raises several important questions. None of which I can pretend to have an answer to.
In the world of the film, the very literal death of the author is part of a dystopian future. But why? Why do we feel uneasy when faced with the possibility of artificial intelligence generating creative fiction? What is it in this notion that confines it to the realm of dystopia? Wouldn’t such computers be to writing what the camera was to painting? If a computer generated a perfect novel, haiku, or even bedtime story, could we enjoy these fictions? Might they be better than our own?
Why does the notion of the machine-generated metaphor terrify me? It could be argued after all, that storytelling began, in the form of myths and legends, as a largely mnemonic practice, a way of storing information about collective histories before the invention of books, libraries, and 2 Terabyte Hard Drives. It’s not just because I’m afraid of computers taking over the world. Something – perhaps it’s the not entirely logical part of my brain (the part, incidentally, that doesn’t mind the odd bit of ‘chick lit’) – demands that I take notice as it screams: there is more to writing than words! -- Isn’t there?
Is writing more than just collection – linguistic curation?
Surely writing and reading fiction demand an understanding of imperfection and uncertainty? Both grammatical and syntactical, as well as metaphorical. Doesn’t writing and reading of fiction ultimately demand the presence of death? Can a machine with perfect memory, with no real foreseeable end point beyond built in obsolescence, a machine which is, to all intents and purposes, immortal, write stories for beings whose time is finite, whose memories are imperfect, and whose own stories have an inevitable end? Can a computer write metaphor and simile meaningful to a human being whose language (whatever the tongue spoken) lacks the clear mathematical certainty of 0s and 1s?
Is Tavernier right? Is mechanised fiction as inhuman and impossible as immortality? In a world where death no longer exists, is writing too necessarily dead?
The French director Bertrand Tavernier (after the work of sci-fi novelist D. G. Compton) imagined such a world in his 1980 film Death Watch – a world where people no longer die of natural causes and where literature is written by machines. There is much critical (and other) literature on the subject of death in the film and its source text. Instead of dealing with death in Death Watch, I focus on the creative, generative aspect – writing, the idea of automated fiction.
Katherine Mortenhoe, the film’s female protagonist, is a ‘programmer’ of romance novels. She is a writer – her sentences, similes, metaphors, and dialogue, however, are programmed. She writes code. Her computer, affectionately known as Harriet, then produces novels that are guaranteed to sell. On one level, the notion of a fiction ‘programmer’ is a clear acerbic comment on the formulaic nature of narrative-building in the genre others have called ‘chick lit’, or a wider comment on the nature of mass-market literature in general.
But more than this, the exploration of the idea of automated fiction raises several important questions. None of which I can pretend to have an answer to.
In the world of the film, the very literal death of the author is part of a dystopian future. But why? Why do we feel uneasy when faced with the possibility of artificial intelligence generating creative fiction? What is it in this notion that confines it to the realm of dystopia? Wouldn’t such computers be to writing what the camera was to painting? If a computer generated a perfect novel, haiku, or even bedtime story, could we enjoy these fictions? Might they be better than our own?
Why does the notion of the machine-generated metaphor terrify me? It could be argued after all, that storytelling began, in the form of myths and legends, as a largely mnemonic practice, a way of storing information about collective histories before the invention of books, libraries, and 2 Terabyte Hard Drives. It’s not just because I’m afraid of computers taking over the world. Something – perhaps it’s the not entirely logical part of my brain (the part, incidentally, that doesn’t mind the odd bit of ‘chick lit’) – demands that I take notice as it screams: there is more to writing than words! -- Isn’t there?
Is writing more than just collection – linguistic curation?
Surely writing and reading fiction demand an understanding of imperfection and uncertainty? Both grammatical and syntactical, as well as metaphorical. Doesn’t writing and reading of fiction ultimately demand the presence of death? Can a machine with perfect memory, with no real foreseeable end point beyond built in obsolescence, a machine which is, to all intents and purposes, immortal, write stories for beings whose time is finite, whose memories are imperfect, and whose own stories have an inevitable end? Can a computer write metaphor and simile meaningful to a human being whose language (whatever the tongue spoken) lacks the clear mathematical certainty of 0s and 1s?
Is Tavernier right? Is mechanised fiction as inhuman and impossible as immortality? In a world where death no longer exists, is writing too necessarily dead?
RSS Feed
