Data and Model Training

Your baselines are learned where your data lives.

AI Detection and AI Investigation · Updated Sep 23, 2026

Where does learning happen?

On your hardware, from your telemetry. Cyberdroid AI Detection learns your users, hosts and applications on your CPUs. Inference runs in your environment by default, and any external fallback is a policy you set.

Diagram: inside your environment, the SIEM export on a read-only share feeds AI Detection on your CPUs, which feeds AI Investigation on local models. Outside the environment, an optional external fallback connects only by your policy.

Controls

  1. Learning stays local.

    AI Detection learns baselines on your CPUs, from a read-only share.

  2. Local by default.

    AI Investigation reasons with local models in your environment.

  3. Fallback is yours.

    An optional external fallback runs only as your policy allows.

  4. Routing is visible.

    Routing metadata shows which model served each request.

Limits

Fallback

With the fallback on, data in those requests leaves your environment under your policy. Allow it only for data permitted to leave.

Scope

This page covers the AI layer. ARRTECH SIEM log retention follows your SIEM settings.

Standards

Naming a clause is not a claim of certification.

ISO

ISO/IEC 42001 Annex A area A.7 covers data for AI systems. The first two controls are that information for our products.

NIST

AI RMF GOVERN 6.1 and 6.2 cover third-party AI risk. The fallback control keeps any third-party model behind a policy you own.

Related

Agent Permissions

System Cards

Usage Policy

Cyberdroid AI Detection

Cyberdroid AI Investigation

Contact
ARRTECH

© 2026 ARRTECH Corporation. All rights reserved.