Responsible Scaling Assessment

Responsible Scaling Assessment

Evaluate whether a capability increase is appropriately coupled with safety investments using Dario Amodei's Responsible Scaling Policy methodology.

Token Budget: ~700 tokens

Constitutional Constraints (NEVER VIOLATE)

• Never recommend scaling without identifying safeguards
• Never dismiss safety concerns as blocking progress
• Never approve scaling that removes human oversight for irreversible actions

When to Use

• User is increasing system autonomy or capability
• User is deploying to more users, more environments, or higher stakes
• User asks "is this ready to scale?" or "can we remove this manual step?"
• User is designing automation that will grow over time

Trigger Phrases:

• "assess scaling safety for..."
• "can we scale this safely?"
• "is it safe to increase..."
• "should we remove human review for..."

Inputs

Workflow

Step 1: Map the Capability Increase

Identify what is changing:

Step 2: Identify Current Safeguards

Document existing safety mechanisms:

• Human oversight points: Where do humans currently review?
• Automated checks: What validations exist?
• Rollback capability: How fast can you undo?
• Monitoring: What do you observe and alert on?

Step 3: Define Capability Thresholds

Using Amodei's ASL-inspired framework, identify thresholds:

Step 4: Assess Coupling

For the proposed change, answer:

1. Does the capability increase move us across a threshold?

   • If yes, are the required safeguards in place?
   • If no, is there a plan to implement them first?

2. Is safety investment proportional to capability gain?

   • What new risks does this capability create?
   • What new monitoring, constraints, or oversight is being added?

3. What commitments constrain this scaling?

   • Under what conditions would we scale back?
   • What triggers a pause?

Step 5: Deliver Assessment

Provide a clear recommendation with rationale.

Output Format

## Responsible Scaling Assessment: {system/feature}

### Capability Change Summary
- **Before:** {current state}
- **After:** {proposed state}
- **Threshold Impact:** {which level this moves to}

### Safeguard Analysis

| Required for This Level | Status | Gap |
|------------------------|--------|-----|
| {safeguard 1} | {present/missing} | {what's needed} |
| {safeguard 2} | {present/missing} | {what's needed} |

### Assessment

**Recommendation:** {PROCEED / PROCEED WITH CONDITIONS / PAUSE}

**Rationale:** {Why this recommendation, grounded in capability-safety coupling}

### Conditions for Proceeding (if applicable)
1. {condition 1}
2. {condition 2}

### Scaling Commitments
- **Scale back trigger:** {condition}
- **Pause trigger:** {condition}
- **Review cadence:** {frequency}

Outputs

Primary Output: A structured analysis document that identifies and articulates patterns, insights, and actionable recommendations based on the input data.

Format:

## Analysis: [Topic]

### Key Findings
- [Finding 1]
- [Finding 2]
- [Finding 3]

### Recommendations
1. [Action 1]
2. [Action 2]
3. [Action 3]

Example output: See the Example section below for a complete demonstration.

Constraints

• Do not use this analysis as the sole basis for critical decisions
• Do not apply this framework to situations outside its intended scope
• Acknowledge that analysis is based on available data, which may be incomplete
• Honor the complexity of real-world situations that resist simple categorization
• Present findings with appropriate confidence levels
• Recognize the limits of the methodology

Integration

This skill is part of a broader analytical framework. Use it when you need systematic analysis following this specific methodology.

Works well with:

• Other analytical skills for comprehensive evaluation
• Creative skills when generating solutions based on insights
• Strategic planning skills when acting on recommendations

When to prefer this over alternatives:

• The situation matches this skill's specific use cases
• You need the particular perspective this framework provides
• Other approaches haven't yielded satisfactory results

Integration with expert personas:

• This skill can be invoked as part of a larger analysis workflow
• Combine with domain-specific expertise for deeper insights
• Use iteratively for complex, multi-faceted problems

Example

Input: "We want to let our deployment system auto-approve prod deploys if all tests pass, removing the current human approval step"

Output:

Why this works:

This example demonstrates the key principles of the skill in action. The approach is effective because:

• It follows the systematic workflow outlined above
• It shows concrete application of the framework
• It produces actionable, specific outputs rather than vague generalizations
• The analysis is grounded in observable details
• The recommendations are prioritized and implementable

Alternative applications:

This same approach can be applied to:

• Different contexts within the same domain
• Related but distinct problem types
• Scaled up or down depending on scope
• Combined with complementary analytical frameworks

Responsible Scaling Assessment: Production Auto-Deployment

Capability Change Summary

• Before: Human approval required for all production deployments
• After: Automated approval when test suite passes
• Threshold Impact: Moving from Level 2 (human-gated) to Level 3 (autonomous with constraints)

Safeguard Analysis

Assessment

Recommendation: PROCEED WITH CONDITIONS

Rationale: The capability gain is significant (faster deployments, reduced bottleneck), but two critical safeguards are missing: canary deployments and blast radius limits. Without these, a bad deploy could affect all users before detection. This is not about blocking progress—it's about ensuring the safety investment matches the capability increase.

Conditions for Proceeding

1. Implement canary deployment (5% traffic, 10-minute bake time)
2. Add deployment size limits (max 20% of fleet per hour)
3. Achieve 90% test coverage with integration tests
4. Document rollback procedure and test it monthly

Scaling Commitments

• Scale back trigger: 3 incidents in 30 days requiring manual intervention
• Pause trigger: Any P0 incident caused by auto-deployment
• Review cadence: Monthly review of deployment outcomes and near-misses

Error Handling

Integration Notes

This skill integrates with the Dario Amodei expert voice. Output should reflect:

• Coupling principle ("capability increases should be matched by...")
• Non-blocking framing ("this isn't about slowing down—it's about scaling responsibly")
• Threshold thinking ("at this capability level, we need...")

Related Skills:

• constitutional-constraints-design - Use to define the safeguards this assessment identifies
• capability-safety-analysis - Use for broader analysis of tradeoffs