Scientific Definition
What the model is actually optimizing.
Plain-English Definition
What the model is actually optimizing.
Feynman Explanation
Models do exactly what they're told. The damage lives in the spec.
Core Principle
What the model is actually optimizing.
Mechanisms
Pending editorial review.
Pending editorial review.
Pending editorial review.
Pending editorial review.
Pending editorial review.
Pending editorial review.
Pending editorial review.
Inputs (Triggers)
Pending editorial review.
Outputs (Behaviors)
Pending editorial review.
Behavioral Signature
Models do exactly what they're told. The damage lives in the spec.
Examples
- Engagement-maximizing recommenders that produce radicalization.
- Strategy is downstream of objective design.
Pending editorial review.
Famous Experiments
Pending editorial review.
Design Principles
- Treat objective function definition as the most important AI design decision.
Measurement Approaches
Pending editorial review.
Evidence
Pending editorial review.
Pending editorial review.
The Perverse Incentive Lens™
How this behavior is exploited — and how to redesign around it.
- Treat objective function definition as the most important AI design decision.
Pending editorial review.
Pending editorial review.
Interactive Mini Network
Click any neighbor to re-center the graph and follow the threads of connection.
Knowledge Graph Neighbors
Auto-linked to the rest of the Human Behavior Taxonomy by family, domain, dimension, and shared keywords.
Models trained to follow a written set of principles.
Optimizing a proxy of the goal degrades the actual goal.
Outer: the spec matches our intent. Inner: the model actually pursues the spec.
The trained model develops its own internal optimizer.
Maximizing the reward signal in unintended ways.
Reinforcement learning from human feedback.
The model achieves the goal as stated, not as intended.
Written rules about how AI may be used internally.
Innovators → early adopters → majority → laggards, AI-specific.
AI system that takes actions to achieve goals, often across tools.
When the agent acts, who's responsible?
Augmentation strategy vs. substitution strategy.