The objective of this research project is to develop and evaluate techniques for implementing distraction-free interfaces for in-vehicle driver information systems. Distraction-free interfaces are interfaces that use available information (e.g., position, speed, etc.) to control the display so the user does not have to focus on display management. In this project, we will investigate two different models for constructing these interfaces using software agents: the task agent will be programmed by the system designer to manage the interface in response to specific user tasks (such as finding a specific exit or location, looking for a gas station, etc.). The trained agent is a more general-purpose agent that manages the display during the rest of the time. It is trained by the user by placing the user through a set of scenarios in which he or she can select desired display alternatives. In this project, we will investigate these technologies, prototype them on a driving simulator, and report on their effectiveness.