In this study, we investigated the effects of context-sensitive distraction warnings on drivers׳ in-car glance behaviors and acceptance. The studied prototype warning application functions on a smart phone. The novelty of the application is its proactive and context-sensitive approach to the adjustment of warning thresholds according to the estimated visual demands of the driving situation ahead. In our study, novice and experienced drivers conducted in-car tasks with a smart phone on a test track with and without the warnings. The application gave a warning if the driver׳s gaze was recognized to remain on the smart phone over a situation-specific threshold time, or if the driver was approaching a high-demand part of the track (an intersection or a tight curve). Glance metrics indicated a significant increasing effect of the warnings on glance time on road while multitasking. The effect varied between 5% and 30% increase depending on the in-car task. A text message reading task was the most visually demanding activity and indicated the greatest effect of the warnings on glance time on road. Driving experience did not have an effect on the efficiency of the warnings. The proposed gaze tracking with current smart phone technology proved to be highly unreliable in varying lighting conditions. However, the findings suggest that location-based proactive distraction warnings of high-demanding driving situations ahead could help all drivers in overcoming the inability to evaluate situational demands while interacting with complex in-car tasks and to place more attention on the road. Furthermore, survey results indicate that it is possible to achieve high levels of trust, perceived usefulness, and acceptance with these kinds of context-sensitive distraction warnings for drivers.