Assessing chatgpt’s potential as an autonomous grant writing tool: a pilot study

Abstract

Abstract Introduction This study evaluates ChatGPT as a potential tool for drafting competitive grant applications to assist researchers in grant-writing. Our hypothesis was that ChatGPT could generate a grant that would be competitive for funding as determined by SAGES Research Committee. Methods ChatGPT was selected for its ability to generate coherent and contextually relevant text. For this study, the SAGES Research Committee grant submission guidelines were used to prompt ChatGPT. The research idea was proposed by the authors. ChatGPT was tasked with autonomously drafting a grant proposal, including generating hypothetical data and a study design. Generated content was submitted without edits to simulate a real-world scenario where AI operates independently. Reviewers were blinded and not aware of AI generated proposals. The grant was randomly assigned to two reviewers with other applications and scored on six criteria and overall level of acceptability. Results The grant received mixed reviews. One rating was “Good,” suggesting it met baseline criteria for quality and relevance. The other rating was “Unacceptable,” citing “feasibility of this trial is highly questionable”. Both reviewers commented the grant was “well-written”. Reviews were statistically different in Potential, Qualifications, and Research environment, where AI performed worse. Conclusions This study demonstrates the potential of ChatGPT as a complementary tool for grant writing. AI was unable to independently produce a proposal that met funding criteria of the SAGES Research Committee. These results suggest ChatGPT has capacity to generate grants approaching professional standards, though not yet consistently at a competitive level by itself.

Ähnliche Arbeiten