GM Super Cruise drivers are the most likely to engage in distracted driving behaviors while using partial-automated driving software when compared to Tesla Autopilot and Nissan ProPILOT drivers, according to a new study by IIHS.
The study was based on a survey of drivers who were asked to self report on which driving activities they had performed and felt safe performing while using partially-automated driving software. All three groups reported a higher likelihood to engage in distracted driving tasks while partial automation systems were turned on. Other drive assist systems were not covered in this survey.
Nissan ProPILOT drivers were statistically the least likely to engage in distracted driving tasks, and Super Cruise drivers were most likely on average, though Tesla Autopilot users were more likely to engage in some tasks than Super Cruise drivers were. Super Cruise drivers were the mostly likely to say they were “comfortable treating their systems as self-driving” (53%, compared to 42% for Autopilot and 12% for ProPILOT) when none of the three systems are actually fully self-driving.
The study asked several questions, including comparisons of whether drivers thought certain activities were safe to do with the system on or off, whether drivers thought they were better at certain activities with the systems turned on, and so on. Here we’ll reproduce a table showing which activities drivers reported doing more often with the system turned on, but for other results you’ll have to click through to the study.
|Percent of drivers who do these things more often with system on:||GM Super Cruise||Nissan ProPILOT||Tesla Autopilot|
|Using phone apps||8||9||23|
|Watching phone videos||3||5||20|
|Talking on cellphone||48||17||33|
|Talking to passengers||47||29||43|
|Hands off wheel (few seconds)||47||15||41|
|Hands off wheel (more than a few seconds)||35||6||46|
|Looking at scenery||63||29||47|
|Looking away from road (more than a few seconds)||58||19||39|
IIHS mentions that this is still early data – it was based on self reporting and is colored by the differing demographics of owners that use these three systems based on the models available that are equipped with them. Tesla and Super Cruise have more male audiences, and Super Cruise tends towards older drivers while Tesla appeals to younger ones (with Nissan having broader appeal). ProPILOT assist users reported using their system more often than Tesla and Super Cruise users.
These demographic reasons could explain why younger and more tech-savvy Autopilot users are more likely to use peripheral devices – phone apps and laptops – than older Super Cruise users.
Most drivers had experienced “attention reminders,” warnings by the system to pay more attention to the road or return their hands to the steering wheel. While some considered these reminders an annoyance, most considered them helpful and said they increase safety of the system. IIHS says this broad consumer acceptance of reminders suggests that distraction reminder systems could be added to more cars without partial automation, as distracted driving is a safety issue regardless of vehicle technology.
Most drivers had also experienced unexpected behavior by the system which required driver intervention, with Autopilot drivers much more likely to experience this unexpected behavior. ProPILOT and Autopilot users were more likely to have had their hands on the wheel when these interventions were needed, and Super Cruise drivers were less likely to have their hands on the wheel (Super Cruise is marketed as a “hands free” system, but the others require occasional steering input).
IIHS cautions drivers to be aware of the limitations of partial driving automated systems and not to exceed those limits. It also calls for more research into driving behaviors while using these systems to better understand whether drivers are using them appropriately and how consumers can be better educated about their capabilities.
The data here is interesting, and shines a little more light on the various sensationalist video clips we’ve seen of Tesla drivers sleeping or reports saying Autopilot has the most ADAS crashes (despite also having the most miles driven on these systems). And on the other hand, it’s also more granular than Tesla’s quarterly Autopilot safety report, which merely does a naive comparison between miles driven on Autopilot vs. overall vehicle safety, without taking into account driving conditions, demographics, age of vehicle, other safety systems, and so on.
It stands to reason that drivers are more likely to engage in these tasks while driving. If you need to grab something from the backseat, look at a sign that’s difficult to read, make a phonecall, etc., then it’s better to do those things with a system backing you up than not. Nobody can expect perfect attention from every driver for every moment, though we can work to minimize driver distraction and fatigue and make sure that there are backups and warnings available to help drivers in moments of inattention.
I’m sure most drivers here who have used these systems have been more likely to engage in some of the tasks listed above. On my recent 2,200 mile roadtrip, I talked through bluetooth and to passengers, drank and ate some snacks, had my hands off the wheel for more than a few seconds, and looked at scenery. When on curvy or crowded roads, I’d be fully engaged, but on open straight roads or when in slow traffic, I did not feel unsafe letting Autopilot manage things while I did something else for a few seconds (or rested my foot).
At the end of the day, responsibility for the car still lies with the driver, and these systems can be used as tools to make driving safer and better or abused in ways that make headlines. More research like this will not only improve how we implement these systems, as IIHS mentions, but will hopefully also result in less sensationalist reporting on their capabilities (or lack thereof).
FTC: We use income earning auto affiliate links. More.