Google’s self-driving division, Waymo, has revealed that its cars receive limited internet access to minimize the window for hackers to penetrate the system.
“Our cars communicate with the outside world only when they need to, so there isn’t a continuous line that’s able to be hacked, going into the car,” said Waymo CEO John Krafcik to the Financial Times. “When we say that our cars are autonomous, it’s not just that there’s not a human driver, but also that there is not a continuous cloud connection to the car.”
Most of the calculations are done inside the car, which is capable of recognizing thousands of objects and understanding their purpose. Krafcik said that the car opens the portal to the internet to send and receive traffic reports, but for the most part remains offline.
Having the car offline for most of the journey and while idle makes it near impossible for hackers to break in from a remote location. It would require knowledge of when the car is sending data to Google’s servers and a way to tap into that encrypted channel.
Waymo cars a lot safer?
This could make Waymo’s cars a lot safer than some other options on the road in 2025, when most vendors expect infrastructure and regulations to be in place for self-driving cars. BMW, Intel, and Mobileye have already referenced 5G wireless service as a key inflection point for self-driving cars and Audi doesn’t see mass adoption before 5G availability.
The need for 5G internet suggests that cars from BMW, Audi, and a few other manufacturers will be always-online, tapping into a massive stream of traffic, weather, and object data. That may reduce the size of some of these cars, which don’t require teraflops of processing power to run efficiently.
While the back-end of the car’s systems may be kept offline, we assume Google will offer a plethora of services and screens for riders to use, which will be connected to the internet. These two systems will most likely be separate, to avoid any security complications.