Call for Papers

We invite high-quality submissions on the verification of AI-generated science. The workshop is nonarchival, and accepted papers will be posted on the workshop website. Submissions are welcome in the following three tracks:

(A) Original Research Track

We invite original studies that develop or apply methods for verifying AI-generated science across all scientific disciplines. Topics include, but are not limited to, learned verifiers, formal methods, surrogate-versus-experiment calibration, uncertainty quantification, active experimental design, and safety-aware deployment.

(B) Position Track

We invite clear, contestable arguments about the epistemics of verification: what constitutes sufficient verification within a domain, when surrogate models can substitute for ground truth, and where current verification practices fall short.

Submissions will be evaluated on the novelty of the argument, its engagement with practice in at least one scientific domain, and the falsifiability of its central claim.

(C) Verifier Systems Track

In the spirit of the NeurIPS Evaluations & Datasets Track, we invite papers describing a deployed verifier. We welcome formal, learned, simulator-based, human-AI hybrid, and consensus verifiers.

Papers should describe how the verifier was constructed, its intended use, its performance and failure modes, and how it can be accessed. Authors must release code and verifier artifacts alongside publication.

Tentative Dates (Anywhere on Earth)

Submission Instructions

All submissions are managed through OpenReview.

The review process is double-blind, so submissions must be anonymized. We welcome original unpublished work, recently published work in scientific journals, and work in progress.

Submissions should be 4-8 pages, with unlimited references and appendices. Appendices are optional, and reviewers are not required to read them. Please use the NeurIPS 2026 LaTeX template; the NeurIPS checklist is not required. Please change the template footnote to “Submitted to/Accepted at/Published in the AI for Science workshop (NeurIPS 2026).”

Contributed talks and best paper awards will be selected based on review scores and discussion among the workshop chairs.