M1.2 Recovering Trust in AV’s

Fully automated vehicles, often also called autonomous vehicles, are something that almost all of the major car manufacturers are working on. Expectations on when the first ones will be driving on public roads can differ depending on who you ask, but according to experts interviewed for this study, they will come. These vehicles are expected to able to drive in a safer manner than humans will ever be able to. Because these fully automated vehicles are expected to improve overall road safety, it is in the interest of all road users that this technology is adopted as soon as possible, when it’s here. Research suggests that trust is the most important factor when it comes to the adoption of automated vehicles, and also that critical situations could harm trust.

In this research project I showed the first evidence that these kinds of situations, if handled correctly, can actually increase trust over time. However, if the user does not understand what happened in such a situation, or doesn’t understand why, or agree with what the vehicle did, this can negatively influence trust. To combat this, I proposed a solution that lets users ask their vehicle for an explanation of its actions. Explanations were given by replaying situations as they were perceived by the vehicle. When comparing trust levels between experiments with and without this functionality, no apparent improvements were to be noted. However, qualitative data did show it was perceived as useful and could help regaining trust.


Figure 1. Slider used for data gathering during experiments