The Value, Benefits, and Concerns of Generative AI-Powered Assistance in Writing

Zhuoyan Li,Chen Liang,Jing Peng,Ming Yin
2024-03-19
Abstract:Recent advances in generative AI technologies like large language models raise both excitement and concerns about the future of human-AI co-creation in writing. To unpack people's attitude towards and experience with generative AI-powered writing assistants, in this paper, we conduct an experiment to understand whether and how much value people attach to AI assistance, and how the incorporation of AI assistance in writing workflows changes people's writing perceptions and performance. Our results suggest that people are willing to forgo financial payments to receive writing assistance from AI, especially if AI can provide direct content generation assistance and the writing task is highly creative. Generative AI-powered assistance is found to offer benefits in increasing people's productivity and confidence in writing. However, direct content generation assistance offered by AI also comes with risks, including decreasing people's sense of accountability and diversity in writing. We conclude by discussing the implications of our findings.
Human-Computer Interaction
What problem does this paper attempt to address?
This paper aims to explore the value, benefits, and potential issues of Generative AI in writing assistance. Specifically, the study evaluates the economic value people are willing to pay for different types of generative AI assistance (such as editing assistance and content generation assistance) through experiments and analyzes how these aids affect people's writing experience and performance. The research results indicate that participants are willing to forgo part of their compensation to receive AI writing assistance, especially in cases where the AI can directly generate content and the task is highly creative. Additionally, the study finds that while generative AI assistance can improve writing efficiency and confidence, it may also lead to some negative effects, such as reducing the author's sense of responsibility and diversity. Therefore, this paper aims to reveal people's true attitudes towards generative AI assistance and its overall impact.