Transparency Log Standard: AI Refusal Logic and GPAI Systemic Risk Compliance

A standardised architecture for the deployment of machine-readable, cryptographically anchored Transparency Logs for General Purpose AI (GPAI) and Model-as-a-Service (MaaS) deployments. Defines structural demarcation criteria for operational telemetry, refusal logic disclosure requirements, a tiered compliance framework (Tier 1-3), and a JSON-LD verification schema (Appendix A) for machine-readable audit manifests. Designed to assist platform operators in demonstrating verifiable alignment with accountability frameworks including Article 12 of the EU AI Act and Section 5 of the FTC Act.