AI‑driven replication audits

Deploy automated AI systems to run standardized replication checks across published social‑science papers, flagging statistical anomalies, undisclosed robustness failures, and likely p‑hacking for human review. These audits would produce machine‑readable provenance reports attached to papers and could be run at scale by journals, funders, or watchdog groups. — If adopted, routine AI audits would shift accountability in research from occasional human replications to continuous machine surveillance, changing incentives for authors, journals, and policymakers who rely on social‑science evidence.