Personas Are Not People: Identity as State, Not Attribute
What it is
This paper argues that using fixed personas for synthetic populations is fundamentally flawed because human behavior is context-dependent rather than a stable attribute. It proposes a 'person-in-context' design that separates stable traits from situational identities to better model real-world behavior.
The problem it solves
Current synthetic population systems assume identity is a stable attribute, failing to capture the significant variation in individual behavior across different contexts.
What you could use it for
- Pretesting surveys with synthetic populations
- Simulating user communities for AI evaluation
- Modeling behavior in groups with high contextual sensitivity (e.g., clinical trials)
Why it matters
Prevents synthetic population evaluations from misrepresenting groups whose behavior is heavily shaped by context, leading to more accurate AI system testing and social simulations.
Read the source ↗Back to the library →
Summary written by an AI model from the source; check the source before you rely on it.