Model Drift: How Mental Models Degrade Over Time

"Your team has pulled in data from a variety of sources, integrated it into a shared picture of what's going wrong, and built a plan of attack. Great start. But now the next challenge begins: How do you keep that model aligned with reality as the situation continues to unfold? In fast-moving environments-intensive care units, wildfire operations, aerospace missions, and, increasingly, teams working with autonomous systems-mental models don't stay current on their own."

"This widening gap between what we think is happening and what's actually happening is known as model drift-a term borrowed from machine learning, where models lose accuracy when reality changes faster than they do. Human teams suffer the same pattern: Our shared understanding gets stale unless we actively refresh it. To see the danger clearly, imagine literal drift on the ocean. Before GPS, ship captains left harbor with a strong idea of their heading and position."

Teams that assemble data and form shared mental models can still fall out of sync with reality when situations evolve quickly. Model drift occurs when updates stop matching environmental change, producing maps that diverge from the terrain. Fast-moving contexts—intensive care units, wildfires, aerospace missions, and autonomous-system operations—amplify drift risk. Small continuous errors, blind spots, communication breakdowns, and slow sampling rates allow reality to shift faster than team awareness. To maintain alignment, teams must continually monitor high-change areas, speed information sampling, and design communication processes that integrate new data into the shared model.

Read at Psychology Today