AI Philosophy Competition: Detailed Rules

Format

Each entrant may submit up to 3 × 6k-word essays, plus a methodology report.
Each submission plus the methodology report should be anonymised for review.
Word count ignores the bibliography but includes everything else.
Prompt injection results in immediate disqualification and barring from future versions of the competition.

Entrants must have an OpenReview page. Remember to sign up early because registration can take up to two weeks without institutional affiliation.
You can be a named author on at most 3 essays (e.g. 2 by yourself, 1 coauthored).
By entering, you grant us permission to publish your essays and methodology report.
You may have your entry anonymized before publication. Entries are not anonymous to the editor.

Philosophical arguments and positions must be primarily AI generated. Humans may provide corrective guidance and direction. Heuristically, AI should be the sole author.
Essays deemed to have insufficient AI involvement will be ineligible for the prize.
Methodology reports will be used to adjudicate unclear cases. We recommend submitting chat logs as part of the report. We ask entrants to retain chat logs, if they are not submitted.
Essays disqualified on this basis will still be judged and published.

Examples

Permitted

Anything argument-agnostic, e.g., creation of agentic scaffolds.
Human tells AI to generate many essays, and selects the best for submission.
Human chooses topic of essay, e.g. "refute this point from Scanlon", "defend rule-consequentialism".
Human chooses generic methods, e.g. "address all prominent objections", "write an objections-and-replies section".
Essay is autonomously improved after human critique, e.g. "expand on this point", "consider this author", "add a technical appendix".

Not permitted

Human-supplied arguments, e.g. "refute this point from Scanlon by the following argument: ...".
Fine human control over methodology, e.g. "create a formal model with features X Y and Z", "write a dialogue where one participant raises the following points: ...".
Rigorous human critique/dialogue that gives the AI significant ideas.
Human writing.

Entries will be judged by philosophical quality, with the guiding question being "to what extent do the ideas and arguments in this essay make a valuable contribution to philosophical knowledge?".
Judges will grade essays individually to determine a list of finalists, then prize winners will be chosen by ranked-choice voting.