In July 2015, two guys remotely hacked a Jeep Cherokee and took over the controls from miles away. The video of the hack went viral, exacerbating fears about the internet of things, connected vehicles, and – by IoT association – autonomous vehicles.
As interest and development in autonomous cars gather steam, fears about hacking and an IoT takeover build, as well. Connected devices inherently trust each other, creating a scenario where a vulnerability in one device is a vulnerability in all devices. A hacker remotely taking control of your home’s autonomous thermostat is one thing, but a hacker remotely taking control of your autonomous vehicle is something else entirely.
But there are still questions that need to be asked and answered. For example: could remote accessibility of self-driving cars be benevolent, or is it always malicious? Also, would the ability to remotely control a vehicle lead to quicker and safer autonomous development?
Surprisingly enough, we can trace the idea of remote controlled automobiles back nearly 100 years.
The first large-scale remote vehicle
In 1925, inventor Francis Houdina sent the first driverless car through the streets of New York City.
The 1926 Chandler was modified with a transmitting antenna that pulled in radio signals from a second car behind it. Houdina took his car up Broadway and then down Fifth Avenue, handling close-knit traffic without incident.
“In the future, the auto tourist, rolling along in strange country, will hear an alarm on his dash sound when he gets off the right road,” reads a 1932 issue of Modern Mechanics and Inventions, “for each great national highway will then have its set of code signals flowing through the telephone or telegraph wires at the side of the roads by means of carrier waves.”
The magazine illustrated Houdina’s idea on a nationwide scale. The cars would still have an antenna on the hood, but they wouldn’t need another car guiding them. The telephone wires (“in addition to their usual duty”) would send out route signals via radio to each car on the road. Accidents would be avoided because the car in front and behind would be following a different radio route along the same telephone wire.
Houdina’s design didn’t entirely remove the driver from the car — a person would still have to drive the car when there wasn’t a signal, or when they needed to alter the route. All the same, it was the first major step toward full autonomy, even though it remained a proof of concept and never reached a mainstream audience.
The case for remote accessibility today
Today’s self-driving vehicle technology, however, will not fall so quietly into the annals of history, says Michael Clamann, senior research scientist at Duke University’s Humans and Autonomy Lab (HAL).
He has a point: Google and other leaders in autonomy (including the director of HAL, Mary Cummings) went in front of Congress in March to guide future federal legislation. Ford says it will have autonomous technology by 2020. There’s even plans for an autonomous Formula 1-style race called Roborace.
Autonomous technology is coming, and it will be able to be remotely controlled. Manufacturers just have to figure how.
“That’s really what the idea behind autonomous cars is,” Clamann tells Inverse. “You’re going to be somehow telling it, with outside coordinates, where an address is, and the car is going to navigate on its own to the location.”
The first example of benevolent remote control that comes to mind for Clamann is in the case of a medical emergency. Say a rider is incapacitated and can’t direct the autonomous car, but a passenger in a passing car sees that he needs to be taken to the nearest hospital. Someone could access the car and reprogram the navigation.
Another argument for remote control is the rideshare economy. Autonomous cars could severely drop the numbers of car ownership, especially in crowded cities. The car-on-demand model could keep car manufacturers alive, and remote connectivity could keep the car-on-demand model alive.
“There are going to be times when you have to have the car empty to pick up a passenger,” Clamann says. “Under those circumstances, someone else may be telling the car: you need to go to X location. Someone, somewhere, has to tell that car where it needs to go, and that could be done with remote control.”
Why the fear of remote hacks won’t slow autonomous car development
“Someone can hack my phone, someone can hack my computer, someone can hack Target,” Clamann says, “that doesn’t stop development of computer systems at these places. Controls can be put into effect, but the fact that people have the ability to hack is not slowing down innovation in these areas, it’s just creating another problem.”
However, in Clamann’s opinion, hackers breaking in and controlling autonomous vehicles remotely is a danger, and should be considered a roadblock in the development of self-driving cars. His concerns echo what Mary Cummings stated in the March meeting in front of Congress — that without proper legislation, autonomous cars will be more danger than they are worth.
John Carlin, U.S. assistant attorney general for national security, stated that he has similar concerns at an automotive conference in April:
“There is no internet-connected system where you can build a wall that’s high enough or deep enough to keep a dedicated nation-state adversary or a sophisticated criminal group out of the system,” Carlin said.
In short: Development of autonomous technology is moving faster than development of security technology, and it’s not stopping any of the manufacturers. It’s the same problem that the entire IoT community is having.
“Don’t get lost in the hype with how exciting IoT is,” Ted Harrington, a cybersecurity expert at Independent Security Evaluators, recently told Inverse, “without balancing it with the risk that comes along with IoT.”
Whether we’re ready or not, there might not be anyone in the driver’s seat of the car next to you sooner than you think.
“As far as autonomous cars go, given the momentum already in the industry, I don’t think it’s looking like it’s going to stall,” Clamann says.
In one potential future, development surrounding remote control of autonomous vehicles could chug along at the same pace, the threat of hackers breaking into our Jeeps lurking around the bend.
In another, security could tighten, and remote connectivity could become a selling point for manufacturers.
Remote control of autonomous vehicles is coming, it’s just a matter of who will be pushing the buttons.