Investigating attitudes of professional writers to GPT text generation AI based creative support tools

This research addresses creative and ethical issues for professional and amateur writers when making AI hybrid or co-created works. We present a mixed-methods study, exploring how these writers used generated texts in a new online creative writing support tool. The study consisted of three creative writing tasks of increasing complexity. Each started with generated text which was then further edited to create a finished work. Feedback on their creative process was gathered as ratings and personal responses, alongside analytics and on-screen activity monitoring. Invited participants were professional and serious amateur writers, including journalists, academics, fiction and non-fiction authors, poets, artists and scientists. By having a wide range of adult respondents we aim to provide in-depth insight into the use of generated text by professional writers. We report on the differences between creative or artistic writers and less creative, analytical or copy writers. Respondent feedback on the custom AI text generator and editor, and their ideas on the integration of text generation with word processors, provide suggestions and paths for future creative support tools. We also gathered insights into the perception of 'fake news' and their relationship to AI generation systems.