AILedger Charter v1.5

Purpose

AILedger makes AI-influenced decisions in regulated industries provable — so customers, regulators, and courts can see what a system did and verify it. The rule we build to is simple: an audit trail is only worth paying for if it gives customers proof they can use — cryptographic provenance of every decision and clear attribution of who, or what, made it. We don't build compliance theater: paperwork that satisfies a regulator while the underlying problem continues helps no one.

Scope: what AILedger does and does not do

AILedger sits at the audit layer of AI-influenced decisions, not the decision layer or the remediation layer:

AILedger facilitates accountability. It does not absolve it. Where a customer needs formal certification, AILedger supports certification partners with evidence; it does not itself certify.

Working as intended

Customers detect bias, drift, and disparate impact in their own systems before regulators or plaintiffs do — early, internally, where issues are cheapest to fix. The populations those systems act on are treated more consistently, and customers carry less regulatory and litigation risk. Regulators and adversarial reviewers treat AILedger output — Decision Events, Integrity Chain verifications, Witnessed Inferences — as evidence. That credibility is the product, and it is what customers pay for.

Failure mode

A product like this fails by becoming theater: audit trails intact and meaningless, thresholds tuned to suppress findings, "we have a compliance tool" standing in for any real change. When that happens the customer ends up more exposed, not less — the appearance of compliance is the first thing a regulator takes apart, and the record AILedger produced becomes the evidence used against them. AILedger is built to avoid this because defensibility is the entire value of the product.

Use cases we decline

AILedger sells across B2B and B2G. The list below is specific use cases — not sectors or company types:

Features we won't build

*Specific standards anchoring detection defaults are listed in STANDARDS.md. The principle of standards-anchoring is the Charter commitment; the specific list is maintained in that document and updates without requiring board approval.

Decisions requiring board review

Exit conditions

Public commitment

This charter is published from day one. Customers, regulators, and the public can hold AILedger accountable to it. Amendments are versioned publicly so changes are visible. A commitment that can be revised quietly is worth nothing to the people relying on it; that is why the revision history is open.

Review cadence

At a minimum, reviewed annually by Board of Directors.

Regulatory context

Updated 2026-05-20, post-Digital Omnibus. The binding obligations on August 2, 2026 are Article 27 FRIA (deployers in credit scoring, insurance pricing, public services, education, and employment), Article 50 transparency, financial-sector high-risk AI, and GPAI provider obligations. Most other Annex III standalone high-risk obligations move to December 2, 2027, and Annex I product-embedded systems to August 2, 2028. This narrows the immediate regulatory surface but does not change AILedger's principle: catch actual harm, not paperwork. The FRIA living-document requirement makes continuous evidence infrastructure more important, not less.

This section is informational. The principles, refused-customer categories, refused-feature categories, and amendment rules above are the binding Charter commitments and are unchanged.

Change log