Automation Complacency: Yet Another Distraction Problem

By Vice Chairman Bruce Landsberg

 The NTSB first issued a recommendation to ban the nonemergency use of portable electronic devices (PEDs) while driving in 2011, and the issue area “Eliminate Distractions” remains on the 2019–2020 Most Wanted List of transportation safety improvements. Web browsing, texting, calling (even hands-free)—all these activities significantly increase the chance of a distracted‑driving crash, which is why we’ve recommended banning driver use of PEDs in all states. Most states have prohibited texting and handheld PED use in some form.

The science is clear: our addiction to PEDs is growing exponentially, placing constant connectivity and convenience above driving responsibly and resulting in tens of thousands of completely preventable, and often tragic, crashes. Driving while distracted by a PED is dangerous and it’s completely preventable. Simply, the decision to drive distracted is dumb.

Distraction by PED is becoming the “old” kind of distraction, as automated and semi-automated vehicles enter the roadways. These new technologies are creating a new and equally menacing kind of distraction: automation complacency. Overreliance on these advanced driver assistance technologies lulls drivers into a false sense of security. They trust in the machine and believe that frees them up to text, e-mail, or watch a video. With automation complacency, human nature asserts itself. We evolve to the idea that we will probably never have to intervene when a computer is doing the driving. The mind creates a rule based on positive prior experience; after so many seamless rides in an automated vehicle, we begin to relax our guard.

A tragic illustration of this growing phenomena is the March 18, 2018, fatal crash in Tempe, Arizona, involving a pedestrian and an Uber vehicle with an experimental automated driving system. The Uber’s safety driver was expected to intervene only if needed, a task that required the driver’s full engagement in and focus on the driving task. Instead, in the half hour prior to colliding with and killing the pedestrian, the driver spent more than a third of her time gazing down at the center console, sometimes for as long as 26.5 seconds. The vehicle’s onboard camera recorded the driver watching streamed content on her cell phone through most of the crash sequence.

Tempe, Arizona crash
NTSB investigators on-scene in Tempe, Arizona, examining the Uber automated test vehicle involved in a March 18, 2018 collision with a pedestrian.

Humans are creatures of habit and this driver had traveled this route more than 20 times in the test vehicle with no incident. Simply put, she was bored. She failed to remain vigilant and succumbed to automation complacency, believing the system would detect pedestrians under all circumstances—even when crossing outside of a crosswalk at night. Our investigation of this fatal crash determined that an attentive human driver would have easily avoided the pedestrian.

If it’s hard to convince drivers to stop multitasking while driving a vehicle that is not equipped with an advanced driver assistance technology, then it’s going to be that much harder to convince drivers to stay alert in a highly automated vehicle. The fact is, there is no commercially available vehicle in the United States that is fully autonomous and doesn’t require the driver’s full attention to the driving task.

The companies testing automated vehicles on public roads, the states where these vehicles are tested, and the National Highway Traffic Safety Administration must work to prevent this emerging form of distraction from increasing and placing roadway users at increased risk, particularly vulnerable users such as bicyclists and pedestrians. Now is the time to get ahead of the problem.

 

 

Leave a comment