Study shows that the more you like your car's personality, the more likely you are to wreck it
Having Homer Simpson read driving instructions to you may end in a big "doh" according to a new study. An international team of researchers has found that the more a person identifies with the voice guidance in their vehicle, the more likely they are to get into an accident.
The test devised by researchers from Michigan State University, Eindhoven University of Technology, and Stanford University began with participants choosing from a catalog of car options, which they rated according to how the vehicle best fit their personalities. They then rated five pre-selected voices based on their perception of the voice's friendliness, human likeness, similarity to them, and intelligence.
A driving instruction simulator was created using an Oculus Rift, a steering wheel, and pedals. The car chosen had a "virtual driving instructor" (voice only) that acted as a "social presence" within the vehicle, giving instructions and directions as the participants drove. The participants were told that their choice of instructor was for research purposes and that the instructor they were ultimately assigned was randomly chosen by the computer, not based on the questionnaire. In reality, the voice given to them was either their highest-rated or lowest-rated.
Drivers who had a voice they strongly connected with in the questionnaire were far more likely to get into an accident during the simulation than were those who did not identify with the voice. The likelihood of a crash further raised if the voice or "personality" of the car was perceived as being similar to their own.
These results suggest that a strong social or personal connection with a car and its virtual voice may hinder safety, and that this factor should be considered when designing autonomous and semi-autonomous guidance systems.
The findings of the paper entitled "KITT, Please Stop Distracting Me: Examining the Effects of Communication in Cars and Social Presence on Safe Driving" will be presented at the upcoming International Communication Association (ICA) conference in Japan.
Source: ICA via Eureka Alert