Pyyan / Compare / New York DIGIT vs EU AI Office vs UK AI Security Institute vs CAISI

New York DIGIT vs EU AI Office vs UK AI Security Institute vs CAISI

4 of 5

Regulators · verified 11 Nov 2026

×NNew York DIGITNew York Department of Financial Servicescurrent
×EEU AI OfficeEuropean Commissioncurrent
×UUK AI Security InstituteUnited Kingdomcurrent
×CAISIUnited Statescurrent
1 slot left
SpecificationNew York DIGITEU AI OfficeUK AI Security InstituteCAISI
SummaryThe RAISE Act's registry and incident desk, open before the law bites.First hard enforcement regime for AI.Frontier model pre-deployment testing.Center for AI Standards and Innovation.
JurisdictionNew York StateEuropean UnionUnited KingdomUnited States
InstrumentRAISE ActEU AI ActSectoralNIST AI RMF
Enforcing since1 Jan 20272 Aug 202620242024
CategoryRegulatorsRegulatorsRegulatorsRegulators
OfficialNew York Department of Financial Services ↗European Commission ↗United Kingdom ↗United States ↗

Highlighted rows are where these differ.

New York DIGIT

  • Large frontier developers, with over $500M in revenue and models trained above 10^26 operations, must register from 1 January 2027
  • Critical safety incidents go to DIGIT within 72 hours, and to a public safety agency within 24 hours if there is imminent risk of death
  • The Attorney General can sue, with penalties up to $1M for a first violation and $3M after
  • Not yet in force: obligations begin in January 2027, and the RAISE Portal opened ahead of that

Best for knowing what New York requires of frontier developers.

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 →

CAISI

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

Full spec sheet →