AI Analysis of George Eliot's Fiction
Chapter Summaries with ChatGPT
In Fall 2023, the computer science students on the George Eliot Archive team, led by graduate student Yuxiang Chen, began experimenting with the public and premium versions of ChatGPT, OpenAI’s large language model (LLM), to learn what it could tell us about Eliot’s works. Results were riddled with errors until we uploaded Eliot’s works chapter by chapter (due to token limits), then prompted the LLM for summaries, character names, and settings analyzing only the provided text. We even asked for “key quotations,” with fascinating results—most predictable but some surprisingly insightful.
We want to remind users that this project is meant not to define or restrict interpretation, but to open a conversation about how an LLM in late 2023 “reads” George Eliot’s works. What new questions can AI prompt humanists to ask?