Milestones

AISYSTEMS is built on a stabilized multi-layer architecture and continues the structured development of its individual layers.

23.02.2026 — Transition to Pilot Deployment Phase

AISYSTEMS has completed the stabilization of its multi-layer architecture (Operational Baseline v4) and is entering a structured pilot deployment phase.

The architecture now clearly separates:

  • the reference layer (AI Core),
  • the methodological authority (AISYSTEMS Audit Framework),
  • the execution layer (AISYSTEMS SENTINEL).

The pilot phase focuses on initial controlled real-world deployments to verify reproducibility, integrity, and deterministic execution under the observe-only principle.

This is not a commercial scaling phase, but a controlled architectural validation in practice while maintaining strict separation of layers.

AI Core v2.0 (LOCKED)

AI Core has been updated to v2.0.0 (LOCKED). Each version is published as an immutable snapshot with cryptographic hash anchoring.

AI Core is a public descriptive ontology. It is not an implementation tool, does not execute audits, and has no regulatory authority.

Current stable version: https://aicore.sk/v2.0/

AISYSTEMS SENTINEL

AISYSTEMS SENTINEL is a deterministic audit execution engine operating in observe-only mode. It executes an AuditPlan and produces cryptographically verifiable technical evidence.

It does not issue decisions. It is an execution layer, not an authority.

The pilot phase aims to confirm that the architecture functions as an institutional model independent of personal authority and capable of generating reproducible technical evidence under real-world conditions.

AISYSTEMS is now operating as a stabilized institutional architecture with controlled pilot deployment.

We produce technical evidence. Decisions remain human.

Current Milestones

16.02.2026 — EMBUILD Phase Initiated

AISYSTEMS entered the implementation phase (HALF architecture).

Focus areas:

  • SENTINEL MVP
  • generation of signed audit artifacts
  • public integrity anchor
  • Base Schema stabilization

The EMBUILD phase follows the institutionalized model and prioritizes technical execution without expanding personal authority or founder dependency.

The focus is on an operational system, not a personality-driven initiative.

16.02.2026 — LOCKED BASELINE v3

The architectural identity of AISYSTEMS v4.1 was formally versioned.

Baseline v3 defines:

  • strict layer separation
  • observe-only audit principles
  • governance boundaries
  • methodological neutrality
  • operational discipline

The architecture was structured as an institutional, role-based system independent of founder presence.

This milestone stabilized the multi-layered architecture as a delegable and long-term sustainable model.

14.02.2026 — LOCKED BASELINE v2

We formally versioned an architectural document defining:

  • the separation of infrastructure, audit, reference layer, and application cores,
  • the observe-only boundaries of the Audit Framework,
  • deterministic principles for evidence generation,
  • neutrality rules, including self-audit.

The architecture is multi-layered and multi-core — individual “cores” can operate independently without compromising the audit methodology or reference layer.

The Audit Framework measures and generates technical evidence. It does not decide.

12.02.2026 — AISYSTEMS s. r. o.

AISYSTEMS was officially registered as AISYSTEMS s. r. o.

What began as an architectural concept of observe-only AI auditing has moved into a formally defined infrastructure phase.

We are building tools, reference layers, and methodological frameworks for reproducible technical evidence.

Reference layers:

  • https://auditumelejinteligencie.sk
  • https://aicore.sk

We produce evidence. Decisions are made by people.

01.02.2026 — Publication of AI Core v1.0 (LOCKED)

We published a public reference ontology: AI Core.

AI Core is a descriptive terminology layer intended for technical, audit, and research contexts. It stabilizes terminology used in observing AI system behavior and distinguishes between:

  • the application layer (where the system runs),
  • the audit layer (where observation takes place),
  • the source layer (the observed model or system).

AI Core does not prescribe implementation and does not provide evaluation or certification.

Version v1.0 is marked as LOCKED — its semantic core is stable. Future versions will be published separately.

Available at: https://aicore.sk