Conveying the Predicted Future to Users: A Case Study of Story Plot PredictionDownload PDF

21 Nov 2022, 19:10 (modified: 14 Jan 2023, 18:03)creativeAIReaders: Everyone
Keywords: story generation
Abstract: Creative writing is hard. Novelists struggle with writer’s block daily. While automatic story generation has advanced recently, it is treated as a “toy task” for pushing artificial intelligence ahead rather than helping people. In this paper, we produce a short description that narrates the predicted plot using existing story generation approaches. We conducted experiments on Amazon Mechanical Turk to examine the quality of the generated story plots in terms of consistency and storiability. Results show that short descriptions produced by GPT-2, a large language model, were rated as the most consistent and storiable among all models. GPT-2’s outputs even beat some random story snippets written by humans. Next, we conducted a preliminary user study using a story continuation task where AMT workers were given access to machine-generated story plots and asked to write a follow-up story. GPT-2 could positively affect the writing process though people favor other baselines more. Our study shed some light on the possibilities of future creative writing supports beyond the scope of completing sentences
0 Replies