EU AI Pact Readiness¶
IINTS includes a research-only MDMP governance checklist inspired by the EU AI Pact and high-risk AI Act readiness themes.
It is not legal advice, not a conformity assessment, and not medical-device certification. It is a stricter internal checklist so IINTS evidence bundles do not overclaim.
Core AI Pact Themes¶
The readiness review checks that a project records evidence for:
- AI governance strategy;
- high-risk system mapping;
- AI literacy.
High-Risk Readiness Themes¶
For stricter research bundles, the review also checks:
- risk management;
- data governance and provenance;
- technical documentation;
- logging and traceability;
- human oversight;
- transparency and limitations;
- accuracy, robustness, and cybersecurity;
- post-deployment monitoring.
Example¶
from iints.mdmp.eu_ai_pact import review_eu_ai_pact_readiness
result = review_eu_ai_pact_readiness({
"mdmp_grade": "research_grade",
"compliance_score": 0.99,
"dataset_fingerprint_sha256": "...",
"contract_fingerprint_sha256": "...",
"row_count": 288,
"governance": {
"ai_governance_strategy": True,
"high_risk_mapping": True,
"ai_literacy": True,
"risk_management": True,
"data_governance_and_provenance": True,
"technical_documentation": True,
"logging_and_traceability": True,
"human_oversight": True,
"transparency_and_limitations": True,
"accuracy_robustness_cybersecurity": True,
"post_deployment_monitoring": True,
},
})
print(result.status)
Interpretation¶
research_ready: the evidence bundle is complete enough for internal research review.needs_review: no hard blocker, but the bundle still needs explanation or cleanup.blocked: do not use the bundle for public AI claims until the missing controls are resolved.
Official Sources¶
- European Commission: AI Pact
- European Commission: AI Act
- European Commission: Navigating the AI Act