# Lab Usage Policies Act as De Facto AI Regulation

Frontier labs will keep writing binding acceptable-use rules on policing, surveillance, tracking and autonomous hardware before legislators act. Their usage policies become a private regulatory layer that government and defense customers must negotiate around, and that layer is a recurring point of friction in public-sector procurement.

- Conviction: 38 / 100 (strengthening)
- Horizon: Emerging (watchlist)
- Tracking since: 2026-10-09T00:00:00.000Z
- Last updated: 2026-10-12T14:07:13.413Z
- Canonical: https://polylog.news/ai/trends/lab-usage-policies-as-de-facto-regulation
- Publisher: Polylog
- Affected regions: United States, Global

## Recent score history

- 2026-10-11: 36
- 2026-10-12: 38

## Recent evidence

- [confirms] Anthropic opens three-tier cyber program giving vetted security teams access to Mythos 5.1 with reduced safeguards (2026-10-12): Anthropic's three-tier Cyber Verification Program decides who gets Mythos 5.1 with reduced safeguards, and at what level, through Anthropic's own vetting. A private lab, not a regulator, is setting the access rules for a dual-use offensive capability.
- [confirms] Anthropic Splits Cyber Access Into Three Tiers, With Grid and Interbank Testing Reviewed Alongside the US Government (2026-10-11): Anthropic now sets three tiers of cyber access, with reduced blocking classifiers on Opus 5.5, Sonnet 5.5 and Mythos 5.1 for vetted defenders. The lab decides who gets offensive-grade capability, and the government takes part as a reviewer rather than the rule-writer.

1 more evidence entry, the full score history, the conviction-driver timeline, and affected assets are for subscribers: https://polylog.news/pricing