ai-safe2-framework

AI SAFE² v3.1 Surface: Research Context: v3.1 Current

Framework Home Research Index Cross-Pillar Governance AISM NEXUS Dashboard

Current framework context: AI SAFE² v3.1. This research note preserves its original publication date, evidence, and historical framework references. Use current v3.1 normative control and profile documents for implementation or conformance decisions.

Research Note: Non-Human Identity (NHI) Sprawl & Governance

ID: RN-2025-002 | Related Control: [P1.T1.4_ADV], [P1.T2.2_ADV] | Status: Verified

🚨 The Threat Vector

Machine-Identity Explosion: As enterprises move to Agentic AI, the volume of Non-Human Identities (Service Accounts, API Keys, Bot Tokens) is growing at 100x the rate of human identities.

🛡️ The AI SAFE² Solution

Standard IAM controls (designed for humans) fail at machine speed. AI SAFE² v2.1 introduces specific NHI Governance Controls:

1. Automated Enumeration [P1.T2.2_ADV]

Agents must be treated as “First-Class Citizens” in Identity Providers (IdP). We mandate automated discovery scripts to map every active agent to its specific permissions map.

2. Ephemeral Credentials [P5.T1.3_ADV]

Static long-lived API keys are prohibited for Tier 3 Agents.

3. Output Hygiene [P1.T1.4_ADV]

Real-time scanning of LLM output streams (using entropy detectors) to ensure an agent does not hallucinate or leak its own configuration secrets to a user.

📚 References


Research navigation

Previous research note Research Index Next research note
Framework Home Cross-Pillar Governance NEXUS Challenge Lab

AI SAFE² v3.1 | Cyber Strategy Institute