Emergent Behavior

Unexpected capabilities or behaviours that arise in large AI models as a result of scale, which were not explicitly programmed or anticipated during development.

Unexpected abilities that appear in AI as it gets bigger, without being explicitly programmed. A language model might suddenly show the ability to do math or reason about logic; surprising even its creators.

Emergent behaviours are a governance risk because they introduce unpredictable capabilities and risks. Your governance framework should include ongoing monitoring for unexpected model behaviours, particularly after model updates or scaling.