US order on frontier-model security
About this eventThe government established a voluntary secure early-access framework and classified evaluation of advanced models' cyber capabilities.
AI governance moved from declarations into operating mode: requirements, audits, reporting, fines. Deadlines and boundaries keep shifting.
About this eventThe government established a voluntary secure early-access framework and classified evaluation of advanced models' cyber capabilities.
About this eventAccess to advanced accelerators remains an export-control instrument: the geography of compute is not set by the market.
About this eventThe next group of rules applies from 2 August 2026, while specific high-risk-system duties follow their own transition schedule.
The AI Act enters into application in stages, so one date does not activate every obligation at once. The important shift is from voluntary principles to legal requirements for documentation, transparency, risk assessment and oversight; each system category must be checked against the Commission's current timeline.
Source: European CommissionAbout this eventThe limited grace period for machine marking and detection of synthetic content ends for systems released before 2 August 2026.
About this eventRequirements for stand-alone high-risk systems in employment, biometrics, justice and border control apply no later than this date.
About this eventEurope plans its own pre-market model evaluation capacity, expected to be operational around 2027.
About this eventThe equivalent of non-proliferation treaties for systems above a capability threshold. Neither a body nor a verification mechanism exists.
Primary material behind this dossier: papers, lab publications and official reports.