When pedestrians prepare to cross a street, they typically rely on eye contact with drivers or interpret subtle vehicle movements to judge whether it is safe to proceed. This simple interaction has worked well for decades, but it breaks down when self-driving cars take control and human drivers are no longer involved. How can pedestrians know they’ve been seen? And how can they anticipate the vehicle’s intentions?

We present and evaluate an external human–machine interface that communicates a vehicle’s perception and intent directly to pedestrians. Results from both simulations and real-world tests show that a front-facing display, positioned between the headlights, significantly improves safety and efficiency. Moreover, it restores the lost social cue of eye contact in traffic while creating a trustworthy and positive user experience.