Levels of Autonomy and the Science of Perception

Trust is subjective and varied. Someone’s level of trust could be wildly different from another’s. And trust between people and technology? That takes it to a whole new level. It’s not surprising then that one common theme among consumers regarding automated vehicles (AV) is “I don’t trust it.”  So we went deep – consumer psychology deep – to find out what does it really mean when people say they don’t trust an AV? 

Trust starts and ends with the perception on feeling safe. 

“For people to feel safe in the car, they need to be able to trust the car completely. But it’s not only about the automation technology… it’s about taking a broader perspective and trusting the future of mobility,” said Nandita Mangal, product lead for autonomous vehicle experience.

After all, it’s no surprise that automation increases the complexity and relationship of trust building between passengers and their vehicles. At Aptiv, we knew in order to understand this topic of trust, we had to approach our research from another angle. There were three key areas of focus: 

• Gather insights on a passenger’s initial reaction with an AV and if it varied based on repeated experiences; essentially transitioning trust from the first ride to seasoned rider. 
• Understand perception of technology limitations and its impact on trust.
• Understand additional techniques passenger use to be aware and orient themselves in an AV, such as monitoring speed changes, looking outside the vehicle and relevant distance. 

Aptiv took an established automotive industry practice – driver simulation testing – and re-imagined it. Instead of testing the driver performance, we focused on ‘testing’ the passengers in automated ride share scenarios, in the back-seat. By putting people in simulated context in a variety of scenarios, we could qualitatively monitor how and when they developed a relationship with the AV to determine how their trust was calibrated, and how long did the process take. In addition, we tracked their eye-gaze and attention patterns throughout a test scenario.


The research involved taking passengers through several tests to gauge their initial level of concerns with an AV current capabilities. It allowed us to glean valuable data that could be used to eliminate superficial perceptions regarding vehicle functionality, and really uncover their worries – preconceived notions and actual attitudes when they thought about mobility and the future.

Participants initially paid a lot of attention to details and tried to establish mental models around  system capability and actual scene reality. Additionally, they understood that they might “learn to ignore” certain system quirks, but there might be a threshold. Intentional and thoughtful design of the in-vehicle experience reassures the reliability of automated technology as well as helps transitioning passenger roles and behaviors inside the vehicle.

We learned passengers have an innate need to orient themselves and become situationally aware just by looking out their window – a behavior which was complemented with internal vehicle interactions. By using this technique, we’ve helped passengers understand the AV system and build trust when they saw the system identifying a potentially dangerous scene, followed by an immediate resolution from the AV. 

Although we cannot test everything in the simulator, we challenged passengers to think beyond the AV and define what they truly wanted in the AV in order to establish trust. We can’t ask someone to imagine the future and expect them to understand its value. We must first create a controlled environment that helps them to fully understand and trust the vehicle. And with the passenger autonomous vehicle simulation… that’s exactly what we’re doing.