days until high-risk enforcement

EU AI Act
High-Risk Systems

Extraterritorial — applies to any company whose AI outputs are used in the EU. €35M or 7% of global revenue for prohibited practices. High-risk obligations begin August 2, 2026.

High-risk AI systems under Annex III

Healthcare
Employment
Education
Essential Services
Law Enforcement
Biometrics

Requirements for high-risk systems

Technical documentation — 9+ required sections covering system design, data, performance, and intended use

Quality management — documented processes for the entire AI lifecycle

Risk management — continuous risk identification, assessment, and mitigation

Data governance — training, validation, and testing data management

Logging — automatic recording of events for traceability

Human oversight — measures enabling effective oversight by natural persons

Penalties

€35M
or 7% of global revenue
Prohibited AI practices
€15M
or 3% of global revenue
Other non-compliance

How GLACIS gets you there

GLACIS generates the continuous technical documentation and evidence trail the EU AI Act requires — automatically, at runtime, with zero data egress.

Technical documentation

Automated generation of required documentation sections from your actual system behavior — not manual PDFs.

Logging & traceability

Every AI decision is cryptographically attested with full provenance — satisfying Article 12 logging requirements automatically.

Risk management

Continuous risk assessment with evidence of controls execution. Maps directly to Article 9 requirements.

Start Free Compliance Assessment

5 minutes. Personalized report. Share with your DPO.