There has been a surge of interest in using Artificial Intelligence (AI) to write psychological science papers, but little research has tested the quality of AI in generating this type of work in psychology. This paper presents a use-case of AI in generating a psychological research paper based on a pre-existing large dataset from the National Health and Nutrition Survey. We aimed to demonstrate to what extent AI can help generate a data-driven hypothesis, perform a literature review, design methodology, complete data analysis, build a discussion, and format a manuscript for journal submission. We discuss the advantages and pitfalls of this approach and propose best practices for using AI to conduct research. Overall, this paper provides an explanation and practical demonstration of the power of existing AI tools in psychological research, while alerting users to their serious attendant risks.
Bouguettaya et al. (Fri,) studied this question.