Regulatory mapping · EU AI Act
Annex IV mapping, out of the box.
Customers building Annex III systems get Article 12 automatic event recording and Annex IV-compatible technical documentation out of the box. We're the infrastructure they audit on — not the regulated AI system itself.
The EU AI Act regulates AI systems by risk tier. If you build an Annex III high-risk system, the obligations fall on you. Ainfera is the infrastructure layer underneath: by running Inferences through Ainfera, the recording and documentation obligations are satisfied by the platform's defaults instead of bespoke compliance plumbing.
Requirement → Ainfera.
Article 12 — automatic recording of events
Every Inference emits a hash-chained sequence of AuditEvents automatically. There is no separate logging step to wire up or forget.
Annex IV §2(g) — record-keeping capabilities
The Audit Chain is the record. It is tamper-evident, timestamped, and verifiable offline by an Auditor.
Annex IV §3 — monitoring and control
Spend policies enforce per-call and daily caps and restrict Providers and Models. Policy checks are themselves recorded as AuditEvents.
Annex IV §2(b) — system design specifications
The five-layer architecture, the Ontology, and ATS v1.0 together document how identity, routing, settlement, and audit behave.
On timing and penalties
The AI Act's general enforcement deadline is 2 August 2026 — about 11 weeks out. Annex III high-risk obligations may push to 2 December 2027 under the May 2026 Digital Omnibus political agreement, but Member State implementations vary and the safe planning posture is 2 August 2026. AI Act penalties run up to €35M or 7% of global turnover for prohibited practices, €15M or 3% for high-risk non-compliance, and €7.5M or 1% for incorrect information.
Related: ATS v1.0 · AAMC v1.0 · Ontology v1.0
This page is informational, not legal advice. Confirm your obligations with qualified counsel.