Model Drift is performance degradation as real-world data shifts. It sits in the Incentives dimension (INC) of the Human Behavior Taxonomy™ as element HBT-INC-0180, within the Workflow family. The core principle: performance degradation as real-world data shifts. In incentive terms, it matters because it changes the payoff people perceive before they choose — which means it can be designed for, or exploited.
Scientific Definition
Performance degradation as real-world data shifts.
Plain-English Definition
Performance degradation as real-world data shifts.
Feynman Explanation
The model that worked last quarter doesn't work this one.
Core Principle
Performance degradation as real-world data shifts.
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
The model that worked last quarter doesn't work this one.
Examples
- Fraud-detection models that degrade as fraud patterns evolve.
- ML operations is half AI strategy.
Pending editorial review.
Original analysis from The Incentives Lab — how this element behaves inside real payoff structures.
Why this element matters to incentive design
When this element shows up in a diagnostic, the instinct is to train people out of it. Training rarely moves it. The mechanism underneath it operates in the Incentives dimension — what makes behavior more or less likely?. You can recognize it in the field by its signature: the model that worked last quarter doesn't work this one. Every element in the Incentives dimension changes the perceived payoff of an action before the action happens, which is exactly where incentive design has leverage.
How it gets exploited
Left undesigned, mL operations is half AI strategy. It is amplified whenever mL operations is half AI strategy. Inside organizations that shows up as mL operations is half AI strategy. The pattern is the same one Goodhart's Law describes: the measurable proxy attracts the effort, and the purpose behind it quietly loses funding.
How the Lab designs around it
The redesign move is to continuous monitoring. Retraining schedules. Drift detection. Treat it as infrastructure. Once you can see it in your own system, most of the argument about culture resolves itself.
Famous Experiments
Pending editorial review.
Design Principles
- Continuous monitoring. Retraining schedules. Drift detection.
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.
- Continuous monitoring. Retraining schedules. Drift detection.
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.
The relationship between inputs and outputs changes.
Underlying data distribution changes over time.
Evaluation suites that no longer reflect real-world conditions.
Train a smaller model to imitate a larger one.
AI system that takes actions to achieve goals, often across tools.
Designing processes from scratch around AI capability.
How much input the model can process at once.
Vector representation of content for similarity and search.
Systematic testing of model quality, safety, and capability.
Models learning from examples in the prompt.
How much delay the user experience tolerates.
Multiple specialized agents coordinating on tasks.
Where Model Drift is cited in the corpus
Essays, field guides, and diagnostics from The Incentives Lab that apply this element.
- Field guideIncentives: definition, types, examples
The parent field guide for this element.
- ReferenceThe laws of incentives
Goodhart, Campbell, and the Cobra Effect.
- EssayAI Agents Inherit Your Incentives
How this element propagates into automated systems.
- ReferenceThe incentive glossary
Definitions for every mental model, bias, and fallacy in the corpus.
- ReferenceThe Periodic Table of Human Behavior
The full 1,267-element map this page belongs to.
Questions about Model Drift
- What is Model Drift?
- Model Drift is performance degradation as real-world data shifts. It sits in the Incentives dimension (INC) of the Human Behavior Taxonomy™ as element HBT-INC-0180, within the Workflow family. The core principle: performance degradation as real-world data shifts. In incentive terms, it matters because it changes the payoff people perceive before they choose — which means it can be designed for, or exploited.
- What is an example of Model Drift?
- ML operations is half AI strategy. The Incentives Lab catalogs everyday, organizational, and historical instances of this element on its Human Behavior Taxonomy™ page (HBT-INC-0180).
- How is Model Drift exploited?
- ML operations is half AI strategy.
- How do you design around Model Drift?
- Continuous monitoring. Retraining schedules. Drift detection.
- Which behavioral dimension does Model Drift belong to?
- Model Drift is classified in the Incentives dimension (INC) of the Human Behavior Taxonomy™, family "Workflow", class "AI-Behavioral Coupling". Its permanent identifier is HBT-INC-0180 and its evidence grade is C.