| Abstract |
While large language models (LLMs) have accelerated scientific production by streamlining writing, coding, and review, the cost of verifying published results remains persistently high. We propose enhancing reproducibility through LLM-assisted verification and demonstrate a working prototype that automatically reproduces statistical research in social science. Quantitative social science is particularly well-suited to automation due to its reliance on standardized statistical models, shared public datasets, and canonical workflows. We introduce a method that iterates LLM-based text interpretation, code generation, execution, and discrepancy analysis, demonstrating its capabilities by reproducing key results from a seminal sociology paper. We also outline deployment scenarios, including pre-submission checks, peer review support, and meta-scientific audits, positioning AI verification as assistive infrastructure that strengthens research integrity. |