Who decides when AI finds something?
A person does. The Cyberdroid SOC Platform, the AI layer of the ARRTECH Security Suite, observes, investigates and communicates, and changes nothing in your estate. Only the suite acts, on your team's rules and approvals.
Controls
See why.
Every finding shows its evidence, its score and the threshold. A score ranks your queue; it is never a verdict.
Touches nothing.
Both products are read-only. Neither can block a host, quarantine a file or change a setting.
Never self-authorizes.
Your policy decides who may ask and who receives the answer. Every decision is recorded.
Stays local.
Both run on your hardware with local models. Any external fallback is a policy you set.
Evolves under control.
No detector alerts until it is measured on your data. Every promotion and rollback is recorded.
Limits
Errors
Any model can be wrong or misled. So no output carries authority, and every conclusion points to evidence a person can check.
Scope
AI Investigation is in early access and requires AI Detection. A third-party SIEM connects through a supported export format.
Claims
These pages describe product controls. They claim no certification, benchmark result or outcome for your environment.
Standards
Naming a clause is not a claim of certification.
NIST
The AI RMF asks for documented human oversight (MAP 3.5), test sets and metrics (MEASURE 2.1) and a way to disengage a system (MANAGE 2.4). Each page names the subcategories it answers.
ISO
ISO/IEC 42001 Annex A includes information for interested parties among its reference controls. These pages are that information for our products.
EU
Article 14 of the EU AI Act describes human oversight: monitor, interpret, disregard or override, and stop. We use it as our benchmark, and Human in Command shows where a person holds each step.
