Four papers
| Paper | What it sets out | SSRN abstract |
|---|---|---|
| From Frameworks to Filings | The AA-1 Agent Governance Attestation standard | 6935198 |
| Living Policy Architecture | The method for compiling written policy into governed rules | 7108198 |
| Policy-coupled Decision Attestation | Attesting each decision against the policy that governed it | 7108538 |
| The DAP Standard 2026.2 | DAP’s classification and taxonomy, used in our regulatory filings | 7377958 |
No papers match this search. Try another term or choose All writing.
Proof, not promises
AA-1
The Agent Governance Attestation standard: how an AI agent’s governance is attested and evidenced. Published on SSRN.
DAAP: the Dartmouth Adversarial Attestation Protocol
Zero-conflict-of-interest evaluation for institutional decision engines. Adversarial by construction, cryptographically verifiable.
- Deterministic replay
- Versioned policy (Merkle tree)
- A zero-knowledge proof per decision
- Public append-only commitment
- Blind adversarial verification
- Verifier learns only pass or fail
At the table where the rules are written
Submissions, October 2026
Comment on listing compute derivatives contracts
Comment on the proposed senior leader regime
Automated-decision brief to the House ETHI committee and Minister Evan Solomon
ISO/IEC 42001 and ISO/IEC 27001, in one integrated audit
Aligned to OSFI E-23 on model risk, effective 1 May 2027, and to ISO/IEC 42005 on AI impact assessment.
We adopt what our buyers are already audited against.
You buy the outcome. We compile the system that delivers it, and the evidence that it did.
Discover a higher standard