Pyyan / Compare / CAISI vs EU AI Office vs UK AI Security Institute vs CAC vs Intl. Network for AI Measurement

CAISI vs EU AI Office vs UK AI Security Institute vs CAC vs Intl. Network for AI Measurement

5 of 5

Regulators · verified 13 Aug 2026

×CAISIUnited Statescurrent
×EEU AI OfficeEuropean Commissioncurrent
×UUK AI Security InstituteUnited Kingdomcurrent
×CACChinacurrent
×MIntl. Network for AI MeasurementMultilateralcurrent
SpecificationCAISIEU AI OfficeUK AI Security InstituteCACIntl. Network for AI Measurement
SummaryCenter for AI Standards and Innovation.First hard enforcement regime for AI.Frontier model pre-deployment testing.Cyberspace Administration of China.Ten-member evaluation consortium.
JurisdictionUnited StatesEuropean UnionUnited KingdomChinaMultilateral
InstrumentNIST AI RMFEU AI ActSectoralCybersecurity LawShared evaluations
Enforcing since20242 Aug 202620241 Jan 20262024
CategoryRegulatorsRegulatorsRegulatorsRegulatorsRegulators
OfficialUnited StatesEuropean CommissionUnited KingdomChinaMultilateral

Highlighted rows are where these differ.

CAISI

  • Renamed from the US AI Safety Institute
  • No federal AI law — state patchwork

Full spec sheet →

EU AI Office

  • Enforcement began 2 August 2026
  • High-risk obligations live the same day
  • EU evaluation capacity expected 2027

Best for .

Full spec sheet →

UK AI Security Institute

  • Outcome-based, pro-innovation model
  • No single UK AI act

Full spec sheet →

CAC

  • AI security review and data localisation
  • Fines and suspensions issued without published reasoning

Full spec sheet →

Intl. Network for AI Measurement

  • Renamed from the Network of AI Safety Institutes
  • Australia, Canada, EU, France, Japan, Kenya, Korea, Singapore, UK, US

Full spec sheet →