arXiv paper proposes ProofTrace for detecting AI involvement in mathematical proofs
Original title: The Gold Rush in AI4Math, II: What are signatures of AI-assisted Proofs?
ProofTrace distinguishes human-written papers from those disclosing substantive AI use with stable performance under AI rewording.
Researchers build five cohorts of 8,281 papers with matched pairs to study AI involvement in proofs. ProofTrace represents each paper as a graph of immediate proof support and extracts interpretable features. The framework reveals qualitative gradients in AI involvement and distinguishes substantive AI use papers with signatures like fewer externally cited results.
- Five cohorts of 8,281 papers with matched pairs
- AI-reworded papers remain close to human papers
- ProofTrace performance stable under AI rewording
- Recurring signatures persist under controls
- Organizations
- OpenAI
From the original article
1 min read at arxiv.orgThe growing use of AI in mathematics raises a question: can we detect AI involvement in proofs, separately from AI assistance with writing? Historical arXiv papers, recent papers disclosing substantive AI use, and manuscripts released by OpenAI provide a rich empirical basis for studying this question.
By Yuanchuan Guo · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.