Use of Artificial Intelligence (AI) by Contributors, Referees, and Editors

Preamble

This policy is written at a time when neither academia as a whole, nor philosophy as a discipline within it, has reached a consensus on the proper use of AI. It is also written at a time when the capacities of AI are changing rapidly. The journal’s stance is therefore precautionary, and in two senses. First, we anticipate that this policy will be liable to change as and when the discipline reaches a more settled opinion on what the role of AI should be in philosophy, and as AI itself changes. Second, this policy works on the assumption that it is less risky to err on the side of having an overly restrictive approach to AI use than it is to err on the side of having an overly permissive approach. Finally, this policy rests on a conception of philosophy as a discipline conducted by persons, for persons, and as a discipline that aims to cultivate understanding rather than merely the generation of truths.

This policy will only be applied to new submissions, and will not be applied retroactively. That said, the Editors-in-Chief reserve the right to act on papers whose submission pre-dates this policy if they judge that the authors have used AI in a way that amounts to misconduct in research.

Policy for Referees and Editors

In the interests of author confidentiality and anonymity, submitted manuscripts (or portions thereof) should never be shared with AI, including Pangram. Referees/editors should never use AI for formulating a report. Referees must also take great care over the general tone of a report, since this sends important signals to the editors. Authors are also liable to pick up on AI-generated text, and would generally prefer to read the referee’s own words. Therefore, they are strongly discouraged from using AI in an effort to improve readability of a report they have already written. If referees/editors choose to use AI for ancillary purposes such as checking quotations in manuscripts, or reminding themselves of relevant positions that may have been published elsewhere, then they must take responsibility for assuring the accuracy of any information they may glean, and they must be mindful of the fallibility of the AI technologies they use.

Policy for Contributors

Authors must make a full declaration of any use of AI in the conceptualization, research, and writing phases of manuscript preparation. This statement will be shared with editors and referees, and should be written in such a way as to preserve the anonymity of the submission. Papers that are subsequently found not to have declared AI use may be subject to retraction. Authors must take responsibility for all content that appears in a submitted manuscript. AI cannot be an author, and cannot be credited as such. The journal strongly discourages the use of AI in substantive roles relating to the overall conceptualization and formulation of a manuscript. Such roles include the generation of arguments, proofs, examples, and so forth; they also include producing summaries of the work of others and of standard positions in the field. Any paper that uses AI in these ways is liable to be desk rejected, and the more an author uses AI for these purposes, the more likely their manuscript will be desk rejected. The journal also strongly discourages use of AI in efforts to improve readability or overall organization of a manuscript; in general, referees and editors would rather see an author’s own phrasing. Again, the more a text is influenced by AI in these respects, the more likely it is to be desk rejected. AI may be used at early stages of the research process—for example, in generating preliminary lists of key publications to read and take into account—but even here authors should be mindful of the fallibility and biases of AI. It may be reasonable to use AI in cases where AI is itself the focus of philosophical research, and it may also be permissible to use AI in certain technical areas (for example, for writing code that is used to build formal models). Even in these kinds of cases, authors should make efforts to seek out and to credit sources that AI may have made use of, and once again authors must take responsibility for all content that appears in their submitted work.

Documenting AI Use

In all cases, an AI statement should appear immediately after the abstract. For example:

The text of this paper was translated to English from German by ChatGPT-5. OpenAI, 19 August 2026, available at chat.openai.com/chat.

Where the AI usage is more complex than translation, full documentation may be included as either an appendix or online-only supplement to the manuscript. In this case, a short AI statement must be provided immediately after the abstract, noting the use of AI and directing readers to the full statement. For example:

The writing of this paper involved the use of AI. A full account of how it was used can be found in the appendix.

This full account must provide readers with the ability to reconstruct the steps taken by the author. Thus, the manuscript must include the following information:

• the specific tool or tools used;
• the date (or date range) when the tools were used;
• for any output, the exact wording of the prompt(s).

Technologies that are not predominately AI tools but which do integrate them must be acknowledged in the same way. For example:

TextTool Pro 3.4 with AI-powered analysis tools was used to…