Researchers: Professor Caroline Bassett, Professor Alan Blackwell, Dr Jo Walton, Dr Anne Alexander
This project introduced early career researchers in the arts and humanities to the generation of texts using neural networks. We explored how ideas about the distinction between ‘fact’, ‘fiction’ and ‘non-fiction’ are shaping the reception of this emerging technology. Our aim was to stimulate deeper critical engagement with machine learning by humanities researchers and to encourage more public debate about the role of AI in culture and society.
We ran a series of online workshops for staff and students at the University of Cambridge and a public event in collaboration with CRASSH.
In collaboration with designer Jana Kleineberg, we produced a booklet exploring the themes of the project which is published below online, and will shortly be available to order in print. In keeping with the project’s goals of investigating Large Language Models as generators of ‘Ghost Fiction’, the booklet aims to provoke debate and reflection through the use of speculative fiction. It was collaboratively written by Caroline Bassett, Anne Alexander, Alan Blackwell and Jo Walton, and includes an imagined study guide and exam paper set in a future world where ‘Automatic Literature’ forms part of the curriculum for undergraduate study.
We gratefully acknowledge the support of the Isaac Newton Trust for this project.
Text Generators Workbook
If you’d like to try out some of the techniques used in the project please see this workbook:
Ghosts, Robots, Automatic Writing
Listen to the Digital Cultures Podcast
We meet the authors of “Ghosts, Robots, Automatic Writing: an AI Level Study Guide”, Dr Anne Alexander and Professor Alan Blackwell and answer the question, can a machine learning algorithm write a Jane Austen novel?
Image: A response to the podcast episode on Automated Writing
Credit: Tom Mclean