Towards a New Psychology of Human-AI Interaction

Press enter or click to view image in full size

Users approach AI systems with social cognition mechanisms optimized for human-human interaction, while AI systems respond through statistical pattern matching optimized for linguistic coherence. Image by author.

“Day by day, what you choose, what you think and what you do is who you become.”

Heraclitus

Introduction

In 1974, Amos Tversky and Daniel Kahneman published their groundbreaking analysis of judgment under uncertainty, revealing that human decision-making follows predictable patterns of bias rather than rational optimization. Their work fundamentally altered our understanding of human cognition by demonstrating systematic deviations from rational choice theory.

The subsequent five decades witnessed psychology’s systematic integration with computing systems, creating what we now recognize as a complete intellectual framework on the verge of replacement.

The 1980s established Human-Computer Interaction (HCI) as a formal discipline, with researchers like Stuart Card, Thomas Moran, and Allen Newell developing cognitive models for interface design. Donald Norman’s work on mental models and Ben Shneiderman’s direct manipulation principles created the theoretical foundation for user-centered design. This era operated under a fundamental assumption: computers were sophisticated tools…