AI is increasingly simulating stable human-like behaviour, often perceived as personality. Stanford HAI research shows these patterns arise from data and optimisation, not true identity. This enables more natural interactions and new simulation use cases, but also raises questions about mistaking imitation for understanding.

