Weekly Brief: US Insurers Pop the Bubble of ADAS Safety Claims

Advanced driver assistance systems got a rude wake-up call last week from the Insurance Institute for Highway Safety.

IIHS is an independent, not-for-profit organization that provides safety ratings for new cars sold in the United States, based on a series of six crash tests. Along with NHTSA’s Stars on Cars program, these safety ratings influence which cars customers feel most comfortable purchasing. IIHS announced last week that it will release a new rating program this year that will evaluate ADAS systems and label them good, acceptable, marginal or poor.

The headline: Not a single ADAS feature on the road today will receive a good rating and some will fail spectacularly.

ADAS systems are built to enhance driver and vehicle safety through features like automated lane changes, adaptive cruise control and lane centering. Many automakers view ADAS as a bridge to the self-driving car future. However, ADAS features are proving more of a danger than a safety enhancement, according to IIHS. Some manufacturers have oversold the capabilities of their systems, thereby prompting drivers to treat the systems as if they can drive the car on their own. In egregious cases, drivers have been documented watching videos or playing games on their phones or even taking naps while speeding down the expressway.

If this sounds like a thinly veiled attack at Tesla, it is. Tesla’s community has earned itself a reputation as technology-loving automotive zealots who believe that their cars can break all the rules while ushering in a new paradigm for driving. On the one hand, they’re right. Tesla has single-handedly changed the game for what cars and trucks will look like in the 21st century. On the other hand, especially when it comes to autonomous tech, they’re often wrong, as has been the case with a number of deadly accidents involving Autopilot in the past five years.

You might recall the fatal crash involving a Tesla Model X that accelerated into a highway crash attenuator in Mountain View, California, in 2018. The National Transportation Safety Board found that Autopilot was engaged at the time, while the driver was playing a video game on his smartphone. I’ve documented a number of similar cases through the years.

What’s noteworthy about IIHS’s announcement is not that it called out Tesla but rather that it broadened the warning shot to the whole industry, including Cadillac Super Cruise, Ford Co-Pilot 360, Volvo Pilot Assist, Mercedes-Benz Active Driver Sensing and others.

“The way many of these systems operate gives people the impression that they’re capable of doing more than they really are,” says IIHS research scientist Alexandra Mueller. “But even when drivers understand the limitations of partial automation, their minds can still wander. As humans, it’s harder for us to remain vigilant when we’re watching and waiting for a problem to occur than it is when we’re doing all the driving ourselves.”

The timing for when IIHS releases its new rating system is uncertain because ongoing supply chain woes in the auto industry have made it more difficult to obtain vehicles for testing but the ratings will come out this year. To earn a good rating, systems will need to monitor both the driver’s gaze and hand position; use multiple types of rapidly escalating alerts to get drivers’ attention when they’ve looked elsewhere or left the steering unattended for too long; and have fail-safe procedures to slow vehicles and keep automation off limits for the remainder of a drive. In addition, automated lane changes must be initiated or confirmed by the driver. Automation features cannot be used with seat belts unfastened or with automatic emergency braking or lane departure prevention/warning disabled.

It’s alarming that not a single ADAS system on the road today meets all of these criteria. It’s equally alarming that none of these criteria are addressed by federal regulations. IIHS factoring them into safety ratings is a start but, ultimately, the federal government needs to find some chutzpah and pass sensible regulations that govern partially autonomous and fully autonomous technologies.