Automated Driving System Abilities Are Often Overestimated: Study

Researchers have found that individuals using self-driving technologies often have inaccurate understandings of system capabilities.

A new study indicates that drivers who use automated driving systems (ADS) are slower to recognize hazards than hands-on drivers, suggesting that growing reliance on automation technology could create new dangers behind the wheel.

The findings were published in the October edition of the journal Applied Ergonomics, examining certain ADS features that manage steering and speed, but still require drivers to monitor the vehicle’s performance and respond to hazards.

In recent years, regulators have launched multiple investigations into automated and self-driving technologies, following two fatal crashes involving Ford’s BlueCruise system, several accidents tied to Tesla’s “Smart Summon” feature, a separate Tesla probe into pedestrian collisions, and a recall of Amazon’s Zoox vehicles following a crash.

As a result, there are growing concerns about the overall safety and reliability of automated driving systems, with research now showing that drivers in vehicles equipped with ADS demonstrate significantly slower hazard recognition, potentially increasing the risk of a crash.

In the new study, researchers from Macquarie University’s Performance and Expertise Research Centre in Australia used a driving simulator to compare participants’ responses to hazards in automated and non-automated modes, asking them to rate their situational awareness.

Led by George Nasser, the researchers found that drivers using automated systems believed they were more aware of their surroundings than those in manual control, yet they actually showed much poorer hazard detection and slower reactions to both expected and unexpected dangers.

The disconnect between drivers’ perceptions and their actual driving performance suggests that people develop inaccurate understandings or mental models of the capabilities of automated driving systems.

“These findings suggest that automated driving may be associated with inaccurate mental models concerning the capabilities of automated vehicles. Therefore, human-centred strategies are required to ensure that drivers are fully cognisant of the capabilities of automated vehicles in recognising and responding to hazards.”

— George Nasser, The mismatch between perceived situation awareness and hazard recognition in automated driving

The findings have led researchers to call for interfaces and visual displays that provide drivers with continuous, easy-to-interpret feedback about system statuses and limitations, helping to build accurate mental models and optimize takeover performance.

Risks of Automated Driving Systems

While automated driving systems are meant to improve road safety, they may also cause challenges, leading to risks for driver and pedestrian safety.

Last month, a jury awarded $329 million in damages to the family of a woman who was killed in a fatal crash involving Tesla’s Autopilot system, which was found to be partially responsible.

Autonomous vehicle manufacturer, Cruise, was also fined $1.5 million by the National Highway Traffic Safety Administration (NHTSA) in 2024, for failing to fully report an accident involving a pedestrian.

In addition, a study from the Insurance Institute for Highway Safety (IIHS) earlier this year examined automatic emergency braking systems, finding that high-visibility, reflective clothing can actually make pedestrians invisible to self-driving cars.

Sign up for more safety and legal news that could affect you or your family.