OECD.AI trustworthy AI crosswalk

MayExecute v2.0 is positioned as an open-source procedural governance gateway that helps AI actors make proposed actions more accountable, reviewable, evidence-aware, safer, and risk-classified before execution.

This crosswalk is prepared by the MayExecute project team for reference and submission purposes. It does not imply review, approval, endorsement, or certification by OECD.AI or the OECD.

OECD.AI trustworthiness themeMayExecute v2.0 supportRelevant EG conditions / v2.0 features
AccountabilityRecords owner, profile, classification basis, reviewer note, and evidence references.Accountability, Reviewability, Human review
Transparency and explainabilityShows each condition, score contribution, risk level, and reason instead of a black-box score.Reviewability, risk classification, record hash
Robustness, security and safetyBlocks or escalates actions when authority, constraints, live context, or proof are missing or invalid.Policy profiles, critical failures, API gate mode
Human-centred values and fairnessEncourages qualified review before AI-supported outputs produce institutional or real-world effects.Verified mandate, valid constraints, accountability
Risk management and governance implementationConverts high-level principles into a lightweight pre-execution gateway with records, hashes, and APIs.Proceed/Review/Block, Low/Medium/High/Critical, Evidence pack hash
Development build — v2.0.0