What was already known
Khardon left characteristic-model completeness complexity open in 1995. Recognizing an individual pseudo-intent was already coNP-complete, but completeness of a proposed canonical list had not been classified.
AI involvement
Mikhail Babin credits GPT-5.6 Pro with the main proof and reduction. Babin checked that proof, wrote the manuscript and explains its reduction mechanism; model autonomy beyond this credit is not established.
The disclosure identifies the model as producer of the main proof, with Babin as checker and author; autonomy is not documented.
Importance
Classifies a longstanding completeness problem, with consequences for related Horn, convex-geometry, formal-concept and database formulations.
Limitations
The model-origin proof is the main closure-representation reduction. Cross-field consequences use established correspondences or separate appendix arguments and are not separate model discoveries. The stated output-polynomial barrier is conditional on P≠NP.
Technical verification
Babin expressly reports checking the complete main proof; this is a source-reported human proof check, not a review reproduced here. The transferred consequences are stated separately.
Formal checking
human check of the model-obtained proof
- Checker
- Mikhail Babin, author
- Scope
- Main coNP-completeness proof for implication–meet-irreducible equality
- Basis
- Reported by the linked checker
Checking a formal statement does not establish that it matches the original problem.
Human understanding
Babin's checked manuscript explains the choice elements, clause gates, auxiliary-cut recurrence and missing meet-irreducible mechanism in §1.4.
Correction and publication history
First report of Babin’s closure-list classification.
Read this published revision