1. Preliminaries

The Philosophy Department is required by the campus to have a policy on AI use by graduate students. It also believes that such a policy, giving relatively concrete guidance, will be welcomed by those graduate students who are uncertain how to proceed but who wish to conform to the Department’s expectations.

The policy is limited in scope to graduate students insofar as they are meeting the requirements of their formal philosophical education. Restrictions on the use of AI that are appropriate for students as part of their formal education may differ from those that are appropriate in other professional contexts (e.g., submitting papers to journals).

The policy only addresses the Department’s view of the proper role of AI in the formal learning activities of graduate students, not in their teaching activities.

2. Permissible AI Uses

Some uses of AI by graduate students are presumptively permissible:

• Using AI as a more convenient search engine.
• Using AI to answer questions of clarification. For example, graduate students who are unclear on the distinction between supervenience and realization may ask an AI tool to explain it to them.
• Using AI to generate unfamiliar objections to their own arguments or conclusions. But if in their evaluated work graduate students address an objection generated by AI, then, to ensure that credit is given where it is due, they must make a reasonable effort to discover the identity of the human originator (if any!) of the objection, so that this originator can be credited in the evaluated work; the AI tool itself may be able to supply this information.
• Using AI to generate feedback on the prose that they themselves have carefully drafted—for example, feedback on grammar, spelling, idiomatic usage, and so on. At the same time, students should take care that the use of AI for this purpose enhances rather than impedes their learning how to write by themselves to the high standard required for doing professional philosophy.

3. Requirements

• All graded student work must include a brief but explicit statement of the role played by AI, if any, in the production of the paper. The statement should identify any tools used, explain how they were used, and affirm that the student reviewed and edited the final work and takes full responsibility for its content. This requirement may be thought of as merely an extension of the longstanding practice in academic writing of explicitly acknowledging the assistance of other people and even of such resources as libraries and archives. If in doubt, err on the side of disclosing more rather than less. More advice may be found here.
• Faculty members teaching classes or advising students may, for their own pedagogical reasons, impose further requirements on their students (for example, requiring them not to complete specific assignments with the aid of AI). Such further requirements could defeat the presumptive permissions above.

4. Recommendations

• Students are advised not to over-trust AI systems, which are not merely fallible but capable of spectacular mistakes that good-faith human sources would never make (e.g., citing non-existent academic papers). Students are therefore advised to verify independently any information obtained from an AI tool, and to keep in mind that they bear sole responsibility for the accuracy of any information they incorporate in their written work.
• Students can only acquire the cognitive skills necessary for reading original philosophy by practicing such reading without assistance. Graduate students are therefore advised not to read AI-generated summaries of assigned readings as a substitute for their own careful reading, and perhaps to read such summaries only after their own careful reading.
• Students should not use any form of AI to generate ideas or to suggest directions of inquiry until after they have made a serious effort to do so without AI (e.g., alone or in conversation with human interlocutors).