#ChatGPT for Research: How to use ChatGPT to Address Reviewer Comments?

Reviewer feedback often turns on a single problem: the manuscript’s arguments don’t yet match the context of the setting being studied. In this session, guidance centers on using ChatGPT to translate reviewer comments into sharper, field-specific reasoning—while warning that any references or claims generated by ChatGPT must be verified before they make it into the paper.

A common reviewer request in the transcript is to explain why findings drawn from corporate CSR literature may not transfer cleanly to higher education institutions, which operate as nonprofit organizations. The approach described starts with a proper response structure: include the manuscript ID, thank the reviewers, then restate each reviewer’s general comments and the specific revision points. When the critique demands more detail—such as how corporate-sector CSR differs from nonprofit CSR—the workflow is to ask ChatGPT for additional argument angles. For example, one suggested distinction is that nonprofit organizations often face stronger stakeholder expectations from donors, volunteers, beneficiaries, and the public, and they rely heavily on trust and funding tied to mission fulfillment. That kind of comparison can help strengthen the manuscript’s logic for why CSR might influence organizational outcomes differently in universities than in profit-seeking firms.

But the transcript draws a hard line around academic integrity and reliability. ChatGPT-generated references “cannot be trusted,” so any citations produced by the model must be checked in Google Scholar or other databases for authenticity. If references are wrong or fabricated, the researcher needs to locate correct sources and replace them. ChatGPT is positioned as an assistant for generating candidate arguments and structure—not as a substitute for reading and building an evidence-based case.

The session also highlights another reviewer concern: the theoretical model links USR (university social responsibility) to university performance, yet the manuscript may have relied on CSR performance effects grounded in consumer behavior literature. The fix is to argue that the “consumer” in higher education is students, and that students’ motivations and responses differ from corporate consumers. That means the literature used for CSR-to-performance claims can’t be lifted wholesale; the manuscript should develop arguments and hypotheses that reflect the unique nonprofit and educational context.

Similarly, the transcript advises that every hypothesis should address the uniqueness of higher education institutions as nonprofits. If the paper proposes relationships like service quality affecting performance, it should clarify how that relationship plays out differently in education compared with profit-based sectors. ChatGPT can help draft these context-specific arguments, but the researcher must still read relevant studies to support them.

Finally, the transcript points to a practical literature review task: when reviewers ask for comparative analysis—such as how USR is adopted in China versus Pakistan—the researcher should prompt ChatGPT for a more specific, country-focused comparison. The output can guide the structure of the literature review, but the underlying claims still require verification through actual scholarly sources. The overall message is clear: use ChatGPT to accelerate thinking and drafting for reviewer responses, then do the scholarly work—reading, checking, and citing—so the final manuscript stands on real evidence.