AJPS AI Policy

From: Dan Reiter and Adam Berinsky, editors-in-chief

We have been discussing amongst ourselves how AJPS should handle the issue of artificial intelligence (AI) in the creation and review of journal submissions. These are important issues we feel we must address. In creating these policies, we had several conversations with other Political Science journal editors. Though there is variance across policies and preferences of journal editors, a common thread is an emphasis on the importance of transparency, of requiring authors and reviewers to disclose if and how AI was used.  There was concern over the lack of reliable tools to detect the use of AI by an author or reviewer, suggesting an emphasis on requiring author disclosure rather than attempting active policing.

With these concerns in mind, we have crafted the following policies regarding the use of AI at AJPS, for both authors and reviewers.

For Authors

“American Journal Political Science requires that manuscript authors must disclose the use of any artificial intelligence tools for work on any element of a submitted manuscript, or any research conducted by the authors to produce the manuscript, for tasks such as copyediting, drafting pre-analysis plans, writing software code, producing mathematical proofs, and others.  This disclosure should be made in the text or footnotes of the manuscript. The text of this disclosure statement must be included in the Author questionnaire at the time of submission.  It is the responsibility of the authors to ensure the validity of any elements that were produced by artificial intelligence.  It is also the responsibility of the authors to ensure that any use of artificial intelligence does not violate ethical guidelines, such as treatment of human subjects.  Authors should avoid using artificial intelligence to write the manuscript or substantial elements of the manuscript, such as the literature review.  Authors must also comply with Wiley’s AI guidelines for researchers.

For Reviewers

“Reviewers may use AI as part of their normal workflow (e.g., finding related papers, copyediting), but reviewers cannot use AI to directly evaluate a paper or write any part of a reviewer report. Reviewers should also comply with Wiley’s AI guidelines for researchers.