White papers
Research on security architecture, evaluation boundaries, and independent evidence. Read the full AI containment proposals and their sources without an email gate.
Research proposal · 2026-10-20 · 6 min read
An Aktoh Cyber design proposal for boundary canaries, decoy endpoints, and independent evidence in AI evaluation environments—with explicit limits and a validation plan.
Read the full paper →Research proposal · 2026-10-20 · 6 min read
A proposed measurement protocol for detection latency, false alarms, path coverage, tamper resistance, and the effect of monitoring on evaluation scores.
Read the full paper →An overview of specialist AI security roles, policy-gated responses, audit records, and the usage model behind the Aktoh fleet.
Architecture overview. Contact Aktoh to ask about the supporting technical material.
Run the whole fleet across unlimited nodes at zero platform fee — you pay only for tokens. Move to a pooled plan when the usage justifies the lower rate.
