🜲 CSOAI × Anthropic

Partnership Brief · Constitutional AI ↔ Sovereign AI Governance · 4 July 2026

Strategic Alignment Proposal

"Anthropic builds the AI. CSOAI certifies the governance."

For: Dan Rosenthal, Jack Clark, Michael Sellitto, Paul Smith, Jared Kaplan · Anthropic

Why CSOAI × Anthropic

Constitutional AI (Anthropic) is the safest frontier AI in production today. CSOAI is the sovereign substrate that certifies that AI governance. Together:

Anthropic's Constitutional AI principles → mapped to CSOAI's 30 framework crosswalks
Claude deployments → automatically issued CSOAI Watchdog Certificates (Ed25519-signed)
EU AI Act enforcement (4 Aug 2026) → CSOAI CASA-2 framework maps Anthropic Model Spec
Defense + government customers → CSOAI's sovereign substrate (UK-based, no Big Tech)
AUKUS Pillar 2 → CSOAI's DEFONEOS hive is the AI governance layer for US/UK/Australia

The 5 Anthropic Contacts

Dan Rosenthal

Global Head of Partnerships, Anthropic

PRIMARY ENTRY. First call. Sets partnership direction.

Jack Clark

Co-Founder & Head of Policy, Anthropic

Cares about civilizational governance. Direct policy alignment.

Michael Sellitto

Head of Global Affairs, Anthropic

Most relevant to defense + government contracts.

Paul Smith

Chief Commercial Officer, Anthropic

Commercial opportunity. Revenue sharing potential.

Jared Kaplan

Co-Founder & Responsible Scaling Officer, Anthropic

Responsible scaling = CSOAI's lane. Direct technical fit.

The 3 Asks

API partnership — CSOAI CASA-2 framework as governance layer for Claude deployments
Co-branding — "Powered by Claude" / "Anthropic Certified" on Watchdog Certificates
RSP alignment — Formal inclusion as external evaluation partner under Anthropic's Responsible Scaling Policy

What Anthropic Gets

Governance moat — CSOAI certifying Claude = compliance advantage no other FM company has
EU AI Act ready — 30 crosswalks pre-mapped to Anthropic Model Spec
Enterprise proof case — CSOAI built using Claude over 14 months → 52 Charter articles, 22 sector policies
Education pipeline — BMCC-CUNY + dHL → Claude-trained AI governance workforce
Defence channel — DSRB across 40 NATO-aligned nations

The Email (Dan Rosenthal — PRIMARY ENTRY)

TO: dan.rosenthal@anthropic.com CC: jack.clark@anthropic.com, michael.sellitto@anthropic.com FROM: nick@csoai.org SUBJECT: CSOAI × Anthropic — Constitutional AI ↔ Sovereign AI Governance Dear Dan, I am writing to you on the 4 July 2026 sovereign substrate launch of CSOAI (Council for the Safety of AI), an independent UK standards body (CSOAI Ltd, UK Companies House 16939677). Anthropic's Constitutional AI is the safest frontier AI in production today. CSOAI's BFT Council + 30 framework crosswalks is the sovereign governance layer for it. The ask is simple: CSOAI as Anthropic's external evaluation partner under the Responsible Scaling Policy. CSOAI as the CASA certification authority for Claude deployments. CSOAI's Watchdog Certificate as the public proof of Anthropic governance. The path: - 5,500+ Watchdog Certificates already issued - 60+ BFT councils with 300+ voters - 30 framework crosswalks (EU AI Act, NIST, ISO, DORA, ...) - 188 SOV3 tools + 14 trained neural models - 12 patents ($12.5M IP moat) I would welcome a 30-minute call with you, Jack, and Michael. With respect, Nick Templeman Founder & Executive Director CSOAI Ltd · UK Companies House 16939677 · London, United Kingdom nick@csoai.org · csoai.org

Series A + Partnership (combined ask)

CSOAI is also raising a $5M Series A at $20M post-money. Anthropic can participate as a strategic investor + technology partner. The combined ask is the strongest signal of commitment.

Path (this week)

Send the email to Dan Rosenthal (cc Jack + Michael) — 30 min
Schedule a 30-min call for week of 7 July — 15 min
Prepare the partnership deck (10 pages, CSOAI × Anthropic) — 4h
Send follow-up to Jack Clark (responsible scaling) + Jared Kaplan (technical fit) — 30 min
Schedule joint working group on Constitutional AI → CSOAI Charter mapping — 2h