Decision Provenance Standard — Frequently Asked Questions
The Decision Provenance Standard™ · v1.3 (reading edition rev. 11) · open standard, CC-BY 4.0
1. What is the Decision Provenance Standard?
It is an open record format: a defined, closed set of building blocks for writing down how a consequential decision was made — by whom, against what information, with what review, and at what moment it was affirmed. It works whether a human or an AI produced the underlying analysis. The goal is to turn "how was this decided?" into a record you can find in thirty seconds and trust, instead of something reconstructed from email and memory.
2. What problem does it solve?
Organizations make consequential decisions every day, increasingly through a mix of people and AI. Later, no one can reliably reconstruct how a given decision was made. As AI authors more of the underlying work, the line of responsibility also blurs. The Standard closes both gaps: it makes decisions affirmable, auditable, and resumable, and it keeps responsibility human as AI does more of the work.
3. Is it legal advice?
No. The Standard is not legal advice and not a regulatory substitute. No attorney-client relationship is created by reading, citing, or adopting it. Any decision with material legal or regulatory consequences should be reviewed by a licensed attorney in the relevant jurisdiction.
4. Does adopting it make me compliant with regulations?
No. The Standard informs frameworks without satisfying them. The records it produces are input that counsel and auditors convert into evidence, certifications, or attestations. A record may be used as input when preparing evidence under frameworks such as NIST AI RMF, ISO/IEC 42001, and the EU AI Act, but it does not itself satisfy any control, requirement, or audit obligation.
5. Who certifies conformance?
No one. There is no certifying body and no certification track. Every conformance level is self-declared by the adopting organization. The Founding Steward governs the text but does not certify, grade, or audit anyone's conformance, and maintains no central registry of conformant organizations.
6. What are Mode 1 and Mode 2?
Every decision is dispatched in one of two modes — answering the question auditors care about most, who actually authored this?
- Mode 1 — Human-Led, AI-Enforced: a human authors the decision; an AI checks it against the Charter's requirements but does not write the substance. AI checks the human.
- Mode 2 — AI-Led, Human-Reviewed: an AI authors the analysis and recommendation; a named human reviews and signs off before any action. The human checks the AI.
There is one more value for the common case of an AI-written passage embedded inside a human-authored document — authorship is judged at the content level, not the container. The enumeration is closed: no third mode, no fuzzy "AI-assisted" label.
7. What is the Charter?
A Charter governs a recurring kind of decision — launch-readiness calls, pricing exceptions, and the like — not a single decision. It states up front, in writing: the single human accountable, the dispatch mode, the schedule of records it will keep, and the triggers that force the decision to be reopened. It moves through five states in one direction — open, mode declared, required fields set, fields completed, closed — and never backward, which makes it consistent by construction rather than by after-the-fact review.
8. What is the disclosure block?
Every AI-authored (Mode 2) artifact carries a small disclosure block, the Standard's own requirement, with five required fields: the declaring authority (the person who prepares the disclosure, named with the organization they act for), the AI system identity, the jurisdiction, the content type, and the timestamp. The block structures the inputs for a disclosure; it does not satisfy any legal obligation. Whether Article 50 of the EU AI Act, or any other law, applies to an artifact, and who carries any duty, is for the deployer to determine. A useful property: when records are later anonymized, four of the five fields survive unchanged — only the declaring authority may be replaced by a placeholder. You can hide who without erasing that an AI authored this, where, of what type, and when.
9. What is the human signature, and why does it matter?
A decision record moves draft → reviewed → affirmed. The load-bearing rule: a record becomes affirmed only when a named human performs an explicit act — a signature — and at that moment it is sealed with a tamper-evident hash. There is no passive promotion: a record never drifts into "approved" because time passed or a box auto-checked. This is what separates the Standard from a logging system, and it is the answer to "won't AI just decide everything?" — no, because a human signature gates every consequential record.
10. What are the conformance levels?
Three cumulative, self-declared levels:
- Level 1 — Charter-Conformant: the Charter is structurally complete.
- Level 2 — Mode-Disambiguated: every record carries its mode, every AI-authored record carries its disclosure block, and an audit finds no silent drift.
- Level 3 — Continuously Auditable: triggers fire on schedule, escalations produce records, disclosures stay current, and the schedule is queryable on demand.
An organization grades itself with a reporter that reads the structural facts. Counsel and auditors read the grade as one input — never as a certification.
11. What is the license?
The Standard is published under the Creative Commons Attribution 4.0 International License (CC-BY 4.0) — free to read, use, extend, and fork under attribution. The reference files (machine-readable schemas, state machines, the signal list, the reporter contract and a test plan, at /standard/v5.0/, reference release 5.2.0) are published separately under the MIT License; they are not the Standard, and the Standard does not depend on them. Etsion Brands Ltd claims "Decision Provenance Standard" as an unregistered trademark (™). The licenses cover the text and files, not the name; permitted uses of the name are set out in Standard §11.1. The Steward asks that no one remove the firewall language or represent the Standard as certified, attested, or regulator-endorsed (Standard §1.5.2).
12. How do I start?
Adoption is voluntary and incremental — the path is Charter-first. Pick one real recurring decision, author a single Charter for it, name the one human accountable, declare whether it dispatches in Mode 1 or Mode 2, begin producing records, and self-declare Conformance Level 1. From there, work upward against the named criteria for Level 2 and Level 3. One Charter, run honestly, is a conformant start.
13. How does it relate to NIST, ISO, and the EU AI Act?
It informs them without satisfying them. The Charter, the decision records, and the conformance grade are an input substrate your AI-governance, compliance, and audit functions can use when preparing their own work — for example, NIST AI RMF Manage 4.1's continuous-documentation expectation, documented information within an ISO/IEC 42001 management system, or disclosure work under EU AI Act Article 50. In every case the Standard discharges no obligation; whether a framework applies, and who carries any duty under it, is for the deployer to determine. The Standard's conformance levels are not ISO/IEC 42001 maturity levels, and there is no defined cross-walk between them.
14. What is — and is not — included in the release?
A small, self-contained set: the core Standard, four Companions (regulatory cross-references, a worked-charter library, implementation guidance, and the diagrams) and an Appendix for governance and references, all under CC-BY 4.0; and the machine-readable reference files (schemas, state machines, the signal list, the reporter contract and a test plan) under the MIT License at /standard/v5.0/, reference release 5.2.0 (not the Standard itself — the Standard does not depend on them).
Jurisdiction assumed: U.S. federal and Delaware (primary); United Kingdom, the European Union, and Israel (named secondaries). For any other jurisdiction, treat every statement here as a hypothesis to verify with local counsel. Not legal advice.
Decision Provenance Standard™ v1.3 — Reading Edition (rev. 11). Open standard under CC-BY 4.0. Not a certified product. Founding Steward: Yohay Etsion; institutional Steward: Etsion Brands Ltd.