Advanced driving assistance systems can do more harm than good, study finds

In newer vehicles, warning systems beep, flash, or vibrate to keep drivers from crashing. Such Advanced Driving Assistance Systems, or ADAS, are standard equipment on most new vehicles. They're meant to make driving more safe by preventing people from drifting out of their lanes or rear-ending other cars.

But are such warnings always helpful? Some of them may be more dangerous than beneficial, suggests new research.

With vehicles now coming equipped with intelligent safety features, researchers have begun to ask whether these systems are changing the behavior of drivers—for better or worse. One group of researchers studied real-world driving data from nearly 200,000 vehicles. The vehicles, sold in 2018 and 2019 by a major automaker, tracked driving behavior using onboard sensors. Some of the cars had ADAS functionality. Others did not.

Ashish Agarwal, a professor at the University of Texas at Austin, worked with Cenying Yang and Prabhudev Konana to determine how these smart alerts affect driving over time. They found that not all alerts work the same way—and some may actually lead to riskier driving.

The researchers looked at two general types of alerts. One was blind spot detection, which warns drivers if there is another vehicle next to them when they want to change lanes. The other was a combination of lane departure and forward collision warnings, which are triggered when a driver drifts out of his or her lane or is about to crash into something.

Blind spot monitoring decreased risky driving behaviors. Cars with it registered 6.76% fewer hard braking events and 9.34% fewer speeding events compared to cars without it. Lane departure and forward collision alerts, on the other hand, had the reverse effect. Drivers with those systems braked hard 5.65% more often and sped 5.34% more often.

Agarwal discusses the reason for these different outcomes is how people think. Psychologists call it dual process theory. It identifies two modes of thinking: System 1 and System 2.

System 1 thinking is automatic and fast. It occurs when you react quickly, like hitting the brakes to avoid an accident. System 2 thinking is reflective and slow. You use it when you plan or think about your mistakes. The problem is that urgent warnings, like forward collision or lane departure warnings, force drivers to respond in the moment. That engages System 1 thinking. "It triggers risk compensation behavior, which inhibits your learning and deteriorates your behavior," Agarwal says.

In contrast, blind spot detection does not demand an instant reaction. It gives drivers time to consider the warning and respond carefully. This is in favor of System 2 thinking, which enables drivers to learn from experience.

In the long term, this learning produces a difference. For each month a driver was equipped with blind spot detection, they sped 0.40% less. But each month of lane departure warnings made drivers speed 0.32% more.

These findings defy the assumption that all warnings are good merely because they alert you to danger. "Our goal here is to give feedback to the auto makers," Agarwal says. "When they're designing these features, they have to be aware that in some cases, they may be making behaviors worse."

The influence of these systems is not limited to braking or speeding in isolation. It even reflects in crash data. Blind spot monitoring caused a decrease in crash rates—down 2.17% through decreased hard braking events, and 3.14% through decreased speeding. This indicates that an appropriately designed non-urgent alert can help drivers avoid accidents over time.

Urgent alerts, however, told a different tale. Lane departure and forward collision systems reduced the safety benefit by 1.71% and 1.66%, respectively, due to higher risky behavior. Even though these systems are meant to prevent collisions, they might make one more likely to happen in the future.

Agarwal and his co-authors say this is because urgent systems may create a false sense of security. Drivers also start to rely on them too much, thinking the car will save them no matter what. It makes people pay less attention to what is going on around them.

This is known as risk compensation. This is where individuals take a bigger risk because they feel they are safe. For example, a helmet might cause a bicycle rider to go faster, even though speed increases the chances of injury. The same is true for certain automobile warnings.

Instead of doing away with warning systems, researchers suggest augmenting them by making them more thoughtful. Warnings can be redesigned to aid learning, not just reaction, Agarwal says. Repeat warnings after a risky maneuver could give drivers time to reflect, for instance. "For learning to take place, you need to be in the System 2 mode," he says. "That is, you learn and your behavior slowly improves."

If car manufacturers keep this in mind, they can help design smarter systems that not only prevent immediate danger, but also help drivers acquire better habits. This could be by giving the driver a breakdown of their driving tendencies or reminding them discreetly once the dangerous situation has passed. Encouraging System 2 thinking allows drivers to become safer in the long term, as opposed to reacting to alerts.

This research may have wider uses too. It shows how smart technology must be built around the way people actually think and behave—not just how we wish they would act. The more that is known about human psychology, the more we can design systems that enable actual improvement.

This study challenges auto makers to consider not just how to alert drivers, but how those alerts shape subsequent behavior. Smart technology has to be more than noise and light displays—it has to make drivers smarter and safer.

The takeaway here is this: Warnings that encourage thought are better than warnings that demand panic. If automakers want safer roads, they'll need to design with the brain in mind.

With smarter cars, it's more important than ever to understand how drivers respond to their digital co-pilots. The future of automobile safety depends on learning—not just responding.

Research findings are available online in the journal Production and Operations Management.

Like these kind of feel good stories? Get The Brighter Side of News' newsletter.