Evaluation & Quality
Concept drift
A change in how useful clues relate to the result an AI is trying to predict.
Example
A purchasing signal that once predicted demand stops doing so after a market change.
Why people use it
It warns that an old connection between information and outcomes may no longer hold.
What you'll hear
“That signal no longer predicts what it used to.”
What this means for you
Monitor outcomes as well as the data arriving at the AI system.
Can you control it?
No
No direct control. This describes a wider issue, concept or result rather than something you can simply switch on or off in a tool.
Common questions
- Can unchanged information become less useful for prediction?
- Yes. The same clue can point to a different result when people's behavior or surrounding conditions change.
- Can it happen even when the information looks familiar?
- Yes. Familiar values can have different meanings after people's behavior or circumstances change.
- Can it be gradual?
- Yes. Relationships can weaken slowly, or change suddenly after a major event.