The European Commission confirmed that the EU AI Act’s requirements for high-risk AI systems β including AI used in law enforcement for evidence evaluation, crime risk profiling, remote biometric identification and predictive policing β will become fully applicable on 2 August 2026 for most systems, with Member States obliged to establish at least one AI regulatory sandbox by the same date. The Digital Omnibus proposal could extend Annex III deadlines to December 2027, but experts warn organisations should treat August 2026 as the binding date. Finland activated national supervision powers in January 2026, becoming the first Member State with fully operational AI Act enforcement.