Longterm Wiki

Responsible Scaling Policies

in-effectInternational
Introduced September 2023Wiki article →

Responsible Scaling Policies (RSPs) are voluntary commitments by AI labs to pause scaling when capability or safety thresholds are crossed. As of December 2025, 20 companies have published policies, though SaferAI grades the three major frameworks 1.9-2.2/5 for specificity.

Introduced
Committee
Floor Vote
Passed
Enacted

Related Topics

Related Pages

Top Related Pages

Risks

Cyberweapons RiskBioweapons Risk

Approaches

Corporate AI Safety ResponsesDangerous Capability EvaluationsAI Governance Coordination TechnologiesAI Safety Intervention Portfolio

Analysis

AI Safety Intervention Effectiveness MatrixIntervention Timing Windows

Key Debates

AI Safety Solution CruxesWhy Alignment Might Be Hard

Policy

Evals-Based Deployment GatesModel Registries

Concepts

Self-Improvement and Recursive EnhancementLong-Horizon Autonomous TasksAgi DevelopmentGovernance Overview

Other

Dario AmodeiElon Musk

Safety Research

Anthropic Core ViewsAI Evaluations

Quick Facts

Introduced
September 2023
Status
Active; 20+ companies participating
Scope
International

Sources

Tags

responsible-scalingvoluntary-commitmentssafety-thresholdsfrontier-labsthird-party-evaluation