THE RSP ARMY

Anthropic's Responsible Scaling Policy
The RSP Army — all 11 characters in formation
Anthropic's Responsible Scaling Policy isn't a single safeguard — it's an entire army of interlocking systems, teams, and commitments working together to ensure AI is developed safely. Click on each character to learn what role they play.
▼ Click a character to learn their role ▼
The Super Soldier
01
The Super Soldier
Claude / The Model
LEADER
The AI at the center of it all.
The Commander
02
The Commander
The Constitution
RULES
The foundational document that governs everything.
The Front Line Guards
03
The Front Line Guards
Constitutional Classifiers
DEFENSE
Automated safety filters screening every interaction.
The War Games Unit
04
The War Games Unit
Red Teaming
OFFENSE
Breaking defenses so bad actors can't.
The Intelligence Unit
05
The Intelligence Unit
Capability Evaluations
INTEL
Measuring exactly what the model can do.
The Situation Room
06
The Situation Room
Risk Reports
STRATEGY
Aggregating intelligence into big-picture risk assessments.
The Oversight Council
07
The Oversight Council
LTBT
JUDGE
Independent authority with the power to intervene.
The Field Marshal
08
The Field Marshal
RSO
OPS
The person who makes sure commitments become action.
The Allied Inspectors
09
The Allied Inspectors
External Reviewers
AUDIT
Third-party auditors who verify safety claims.
The Civilian Militia
10
The Civilian Militia
Bug Bounties
BOUNTY
The public helping find what internal teams missed.
The Whistleblower Channel
11
The Whistleblower Channel
Noncompliance Reporting
COVERT
Protected channels for raising safety concerns from within.