AI Policy

Education Policy Analysis Archives (EPAA/AAPE) Policy on Artificial Intelligence (AI)

Author(s)

Authorship

• AI cannot be an Author. Generative AI tools (e.g., ChatGPT, Large Language Models) do not meet the criteria for authorship because they cannot take legal or ethical responsibility for the work, manage copyright agreements, or assert the absence of conflicts of interest.
• Human authors are solely responsible and accountable for the accuracy, integrity, and originality of the entire manuscript.

Manuscript Preparation

Permissible Use: Author(s) may use AI tools for:

• Language editing or grammar assistance
• Improving clarity or readability
• Translation support
• Alt text for images
• Checks on qualitative findings or producing initial code, but author(s) are responsible for verifying all content.

Restricted Use: AI is not permitted for the following:

• Authorship

According to COPE, WAME, JAMA, the US copyright office, etc., authorship requires a "legal entity" status to enter into copyright agreements and manage conflicts of interest. Because AI cannot take ethical or legal responsibility for research results, listing it as a co-author is a violation of publication ethics.

• Compiling or drafting the literature review
• Analyzing data
• Generating data, references, or citations
• Generating images or figures
• Conducting peer review

Important: In the event of ambiguity or uncertainty regarding AI usage, author(s) should contact their section editor, or the managing editor, or lead editor.

 

Disclosure Requirements

• Any use of generative AI must be disclosed at the time of first submission or revision.
• How to Disclose: Two separate declarations are required, including to (1) the lead editor; and (2) in the manuscript itself.
  • Immediately after receiving the submission acknowledgement email, the author(s) must submit a Disclosure of AI Use using the Add Discussion feature. On the Workflow tab (tip menu) under Submission, go to the Pre-Review Discussions section and add a discussion labelled DisClosure of AI Use.

Manuscripts will not be reviewed unless the author(s) provide the AI disclosure.

• Authors(s) must include a declaration statement within the text itself (highlighted and clearly marked) so that reviewers are aware when assessing the manuscript.

• In both declarations, author(s) must provide:
  • The names and versions of the AI tools used.
  • The specific task the tools performed (e.g., "summarizing data," "drafting portions of the introduction", “language proofing”).

 

Editors and Reviewers

• Reviewers and editors must not upload any part of a manuscript into a generative AI tool. Uploading manuscripts violates authors’ confidentiality and proprietary rights.
• Disclosure: If a reviewer uses AI to edit the language of their report, they must disclose this use to the editor managing the review process for the manuscript.
• Editors and reviewers are fully responsible and accountable for the content, accuracy, and integrity of their reports.

 

Protocol for Misuse and Investigation/Appeals Process

Queries to author(s) concerning their use of AI is not an accusation but an investigation. Investigations of potential misuse of AI tools will be handled on a case-by-case basis. If an editor or reviewer identifies questionable AI use by the author)s) in their manuscript, they first notify the lead editor or managing editor. The lead editor will contact the author(s), stating the issue and ask the author(s) to explain their use of AI.

• If the author(s) do not respond or do not address the questions raised, the manuscript will not move forward in the screening or review process and be promptly archived.
• If the author(s) provide a sufficient explanation concerning their AI use and/or agree to follow the Journal's recommendations to modify, remove, or replace the text, data, or graphic presentations in question, the screening or review process may proceed.