Blueprint ranges
I · Foundations of AI governance
I.AWhat AI is and why it needs governance4-6
I.BOrganizational expectations and roles5-7
I.CPolicies and procedures across the life cycle6-8
II · Laws, standards and frameworks
II.AExisting data privacy laws applied to AI4-6
II.BOther existing laws (IP, nondiscrimination, consumer, product)4-6
II.CAI-specific laws (EU AI Act, Korea, US federal and state)6-8
II.DStandards and tools (OECD, NIST, ISO)3-5
III · Governing AI development
III.AGovern design and build6-8
III.BGovern data in training and testing6-8
III.CGovern release, monitoring, maintenance8-10
IV · Governing AI deployment and use
IV.AFactors and risks in the decision to deploy6-8
IV.BAssess the AI system (impact, vendor, proprietary)5-7
IV.CGovern deployment and use9-11
The exam
Items100
Scored85
Time3 hours
Passing scaled score300
Fee (member / non-member)$649 / $799
EU AI Act dates
Prohibitions (Art. 5)2 Feb 2025
GPAI obligations2 Aug 2025
Transparency (Art. 50)2 Aug 2026
Annex III high-risk, as written2 Aug 2026
Annex III high-risk, after Omnibus2 Dec 2027
Annex I high-risk, as written2 Aug 2027
Annex I high-risk, after Omnibus2 Aug 2028
Digital Omnibus in forceJuly 2026
Omnibus ban on non-consensual intimate imageryDec 2026
Not moved by OmnibusProhibitions, GPAI, Art. 50
EU AI Act thresholds and fines
GPAI systemic-risk presumption10^25 FLOPs
Prohibited practices€35M or 7%
Most other obligations€15M or 3%
Misleading information to authorities€7.5M or 1%
GPAI providers (Art. 101)€15M or 3%
RuleHigher applies; lower for SMEs
South Korea AI Basic Act
In force22 Jan 2026
Compute threshold10^26 FLOPs
Max fineKRW 30M
Impact assessmentShall endeavor (best efforts)
Domestic rep. triggersKRW 1T revenue, KRW 10B AI revenue, or 1M daily users
Record keeping5 years
US state and federal
Colorado SB 26-189 effective1 Jan 2027
Colorado replacesSB 24-205 (repealed May 2026)
Colorado adverse-outcome window30 days
Colorado enforcementAttorney General only
Texas TRAIGA effective1 Jan 2026
OMB memorandaM-25-21, M-25-22 (Apr 2025)
Four-fifths rule0.8 selection-rate ratio
Frameworks and standards
NIST AI RMF functionsGOVERN, MAP, MEASURE, MANAGE
NIST trustworthy characteristics7
NIST AI 600-1 GenAI risks12
ISO/IEC 42001Certifiable AIMS, 38 Annex A controls
ISO/IEC 42005Impact assessment guidance, not certifiable
ISO/IEC 22989Concepts and terminology
OECD AI Principles5 (updated May 2024)
GDPR breach notice72 hours