The researchers will present their research next week at the Black Hat cybersecurity conference in Las Vegas.
Christian Werling, one of the three students at Technische Universität Berlin who conducted the research along with another independent researcher, said that their attack requires physical access to the car, but that’s exactly the scenario where their jailbreak would be useful.
“We are not the evil outsider, but we’re actually the insider, we own the car,” Werling told TechCrunch in an interview ahead of the conference. “And we don’t want to pay these $300 for the rear heated seats.”
The technique they used to jailbreak the Tesla is called voltage glitching. Werling explained that what they did was “fiddle around” with the supply voltage of the AMD processor that runs the infotainment system.
“If we do it at the right moment, we can trick the CPU into doing something else. It has a hiccup, skips an instruction and accepts our manipulated code. That’s basically what we do in a nutshell,” he said.
With the same technique, the researchers said they were also able to extract the encryption key used to authenticate the car to Tesla’s network. In theory, this would open the door for a series of other attacks, but the researchers said they still have to explore the possibilities in this scenario.
The researchers said they were also able to extract personal information from the car such as contacts, recent calendar appointments, call logs, locations the car visited, Wi-Fi passwords and session tokens from email accounts, among others. This is data that could be attractive to people who don’t own that particular car, but still have physical access to it.
Mitigating the hardware-based attack that the researchers achieved is not simple. In fact, the researchers said, Tesla would have to replace the hardware in question.
Tesla did not respond to a request for comment.
So they install heated seats and then make you pay to unlock them?
That seems... not cost-effective
Actually, on that scale, it probably is more cost effective. They don't need separate factory lines, or to pump out cars with all sorts of different combinations of options. It takes better advantage of the economy of scale.
Why not just have the seats as a base model feature?
I can understand having one wiring harnesses, and having two types of seats, but otherwise you'd need to charge twice the cost of the feature to break even.
Twice of the much lower cost, due to economies of scale … which may/should still be significantly less than building two - and more, for more features - factory lines and risk not selling a car for a longer time, since it doesn’t have the right feature combination.