Responsible AI Statement
VeriSigil AI was founded on the principle that governance should be infrastructure — not afterthought. This statement describes how we apply that principle to our own operations and to the tools we build.
Our Core Principles
Human Sovereignty is Non-Negotiable
We build systems where humans remain the final authority on consequential decisions. Eight categories of decisions are constitutionally reserved for human authority in our platform — employment, medical treatment, legal prosecution, military action, and others. This is architectural, not configurable.
Governance Before Execution
Our constitutional model requires that AI systems satisfy admissibility checks before taking consequential actions — not after. We do not build monitoring tools that observe harm after it occurs. We build execution gates that prevent unauthorised action before it happens.
Transparency Through Evidence
Every governance decision made through VeriSigil infrastructure generates a cryptographically sealed evidence record. These records are independently verifiable offline — they do not require trusting VeriSigil. We believe governance evidence should outlive the platform that generated it.
Honest Claims
We align with EU AI Act, NIST AI RMF, and ISO 42001 governance principles. We do not claim official regulatory certification. Our protocol attestation is independent validation of governance behaviour — not government accreditation. We say this clearly and consistently.
No Weapons, No Harm Amplification
VeriSigil AI does not provide services designed to amplify lethal autonomous systems, surveillance without consent, or systems designed to cause harm. We reserve the right to refuse access to any use case that conflicts with this principle.
Accountability Continuity
We build governance infrastructure designed to function even if VeriSigil ceases to operate. Evidence bundles are offline-verifiable. The VSL grammar is open. The Constitutional Charter is published under DOI. Your governance record does not depend on our survival.
About This Statement
This statement reflects our current principles and practices. We will update it as the platform and company evolve. We welcome feedback from researchers, regulators, and enterprise partners.
Contact: info@verisigilai.com
Published Standards
- VGS-ELI Specification: doi.org/10.5281/zenodo.20451306
- VSL Grammar (govlang.ebnf): CC-BY-4.0 — open standard
- SDK: MIT licence — github.com/raheem-verisigil/verisigil-api