How much do people trust automated vehicles (AVs), and do trust levels change with the amount of exposure one has to them?
A new UC Berkeley SafeTREC study surveyed 811 San Francisco residents who have had real-world exposure to AVs — 376 robotaxi passengers and 435 non-passengers who are pedestrians, bicyclists, or drivers in the city — to determine how much they trust AVs and what concerns they had about them, if any. This data was then analyzed to answer the following questions:
1. Do trust levels among AV-exposed residents exceed national baselines, and do non-riders also show elevated trust?
2. Does trust vary by interaction mode (passenger, pedestrian, cyclist, driver)?
3. Among passengers, does cumulative experience predict higher trust? Among non-passengers, are close calls associated with lower trust?
4. Do concern patterns differ by mode? Specifically, do technology-related concerns attenuate with passenger experience, and do human-behavior concerns persist?
The report reveals that 61% of all survey participants trust AVs, which is 4.7 times more than the national rate of 13%. Trust rates varied widely depending on survey participants' exposure to AVs: 82% of AV passengers trust them, while only 45% of drivers and 43.2% of pedestrians and cyclists do. This may suggest that ambient exposure to AVs can increase trust levels in AVs, although further studies would need to be conducted to rule out other factors.
Safety ratings of AVs also correlate with the amount of times one has been inside of one; first-time passengers of AVs trust AVs more than non-passengers, and repeat riders report feeling safer than first-time riders. Among non-passengers, close calls with AVs were strongly associated with lower trust levels. 18% of pedestrians and cyclists with close calls trust AVs compared to 49.4% without, and 26.7% of drivers with close calls trust AVs compared to 51.1% without.
For more information about the study and access the full report, visit Science Direct's page for "Trust in the wild: Multi-modal safety perceptions in naturalistic AV deployment."
Funding for this study was provided by a grant from the California Office of Traffic Safety, through the National Highway Traffic Safety Administration.