The Routledge Handbook of AI and Literature, edited by Genevieve Lively and Will Slocombe, will be out in December. I've got a chapter in it, offering a thunderously gloomy assessment of Kai-Fu Lee and Chen Qiufan's AI 2041: Ten Visions of Our Future. I don't mean to be like this, I obviously fell into a cauldron of Adorno when I was a baby.
The chapter also offers the concept of "critical design fiction." Chen’s intro to AI 2041 suggests that science fiction has the ‘capacity to serve as a warning’ but also that ‘every future we wish to create, we must first learn to imagine.’ Critical design fiction is my attempt to somewhat formalise and test this idea. Critical design fiction would be fiction which:
- adjusts the probabilities that what it represents will occur,
- if it had been written differently, it would have adjusted those probabilities differently,
- decreases the probability of something it represents occuring, OR could have decreased its probability if it had been written differently.
By ‘what it represents’, I'm talking about the general type of events or states of affairs, rather than the specific details of what the characters do or experience. So although it brings some precision, there is plenty of room for interpretation and contention. Similarly, I haven't tried to formalise the difference between 'the same story written differently' and a 'different story.'
Why am I so interested in 'decreasing the probability' of things? Critical design fiction is in part a response to science fiction's 'Don't Build the Torment Nexus' problem.
The chapter, by the way, got far too long, and so half of it budded off into a separate essay: 'Machine Learning in Contemporary Science Fiction,' over at The SFRA Review.
Massive thank-you to Genevieve and Will for their thoughtful and patient edits. The AI and Literature handbook has a formidable academic publisher price, so this is one to ask your library to order. If you are desperate to read something in it and can't get access, let me know and I'll see if there's anything I can do.
No comments:
Post a Comment