Assessing 'Skill Decay' in Autonomous Driving
Aug. 19, 2020—Autonomous driving is often talked about with a positive connotation because it is regarded as the future of transportation. But with all of its ease and convenience it gives way to an entirely different issue—skill decay.
In a virtual webinar hosted last week by Partners for Automated Vehicle Education, Michael Clamann, a human factors engineer at the University of North Carolina’s Highway Research Safety Center, warned against the loss of manual skills as a direct result of automation.
It is one thing to note what the limitations of automation realistically are, it is an entirely different discussion to determine what the limits should be, he says.
“Machines have quick response times and are able to apply force in a consistent and efficient manner,” Clamann says. But that begs the question, which jobs do you assign to the computer and which jobs do you give to the human?
Michael Nees, an associate professor at Lafayette College says in automation, the human is not incurring a workload but instead they are monitoring the systems and looking for rare and infrequent hiccups.
“The issue with that is, vigilance and monitoring are not the best tasks to give to a human,” Nees says.
Duke University Professor of Robotics and Artificial Intelligence, Missy Cummings, says driver distraction coupled with a semi-capable autonomous vehicle is a dangerous combination.
“If you leave humans to babysit highly automated systems, you can be sure they’re going to get distracted,” she says. “It’s almost painful to sit and watch automated systems and wait for something to happen.”
Clamann says that on average, after 20 minutes of monitoring these systems, people lose interest. It is not only everyday drivers who cannot stay focused on the task, but even AV operators running tests at the vehicle facilities.
Over time, our brains have been re-wired to seek information constantly. With evolving technology comes convenient information and when everything is one touchscreen away, it makes it incredibly difficult for the general public to focus on one thing at a time.
“That desire to constantly seek information is pulling people away from the task of monitoring,” says Cummings.
There is no quick fix to this issue. In fact, Clamann says there is no solution at all.
“You cannot train out human nature. You are still going to look elsewhere for distractions and attempt to find that ‘sweet spot’ of mental workload, despite the importance of the task at hand.”
Nees suggests a new way of reframing how automation functions across all spectrums in order to put the focus back on the human. He says a human-centered design approach rather than an automated-center design approach could be the key.
“Most of the discourse surrounding technology has begun with the assumption that we will engineer the ‘annoying human’ out of the system,” says Nees. “But the human is not an inconvenience, rather the system should be taking over what the human cannot do.”
Automation is a brilliant, albeit new, technology. But, when there is a computer system at the helm, drivers tend to overtrust its capabilities.
“We are all human and we all become complacent. If the system is doing just enough, it will lull us into a false sense of safety,” says Cummings. That is why there is still work to be done to bridge the gap between human and computerized capabilities.
“We have standardized air bags and seat belts and next, we’ll standardize automation,” she reassures.