What is AI neutralization?
AI neutralization is a controlled defensive response to an AI system that is behaving outside its authority. It progresses through detect, classify, limit, quarantine, revoke authority, disable the controlled agent, safe decommission, and evidence plus human review. It applies only to systems and resources the operator is authorized to administer — never to third-party systems.
- Graduated: least disruptive effective step first.
- Human decision point before every irreversible step.
- Out-of-organisation targets are rejected (HTTP 403).
Frequently asked questions
Is AI neutralization an offensive capability?
No. It is authorized containment of your own systems: limiting, quarantining, revoking, disabling and decommissioning agents you administer.
Can a neutralized agent be restored?
Yes, until decommission. Limits, quarantine and revocation can be reversed after human review.