Part I: System maps and threat records

A system record shows what exists and who controls it; a threat record adds a specific attacker, failure path, control, evidence, and remaining risk.

An employee support assistant supplies one fixed system. Its system record identifies people, components, data, operating modes, permissions, possible state changes, and available evidence. The threat record keeps that system fixed while adding an unwanted outcome, attack path, local control, test, decision owner, and review trigger.

Inside the organization boundary, an employee session reaches an AI application, document store, action service, and business records, while an external model provider sits in a separate dashed boundary. An external attacker can send a hostile request or document. Below the system, a Chapter 1 system record feeds a Chapter 2 threat record, whose local control points back to the action service.
Figure 1: Chapter 1 records the fixed system, including its people, components, data paths, checks, protected assets, assumptions, and evidence gaps. Chapter 2 keeps that system fixed while adding an unwanted outcome, attacker access, an attack path, a local control, test evidence, remaining risk, a decision owner, and a review trigger. The action service remains separate from the model, so a generated proposal cannot change business records without current authorization.

Chapters in this part

  • Mapping the AI system: A system map records the people, components, data, permissions, and evidence that a security decision about an AI assistant depends on.
  • Threat modeling and controls: A threat record connects an attacker’s access to a possible failure, the controls expected to prevent it, and the evidence needed to test them.