Let me start by saying I am confident that the connected, self-driving car will happen sooner than any of us can imagine. Having said that, let me also say that success is hidden behind some freakishly huge obstacles. There is an intelligence disconnect between what connected cars are currently capable of and what will be required for an automated network of self-driving vehicles.
The gravity of the situation demands an altogether higher order of security than we have at the moment. Consider the state of our current connectivity.
Given the remote control and locate functions of applications like GM’s OnStar, is there any reason for vehicles to be stolen and lost? The answer is Yes. You don’t have to look further than the fact that these systems can be hacked or disabled. Hacking for the purposes of theft is one thing; hacking or disabling an automated vehicle traveling at city speeds is another altogether.
If technology companies themselves can be hacked (see Sony), broadly distributed utilities like a self-driving vehicle network are certainly at risk. Firstly, their size makes them a prized target for both benign and malicious hackers. On the same token, their size makes them exponentially more difficult to secure.
Every port, communication protocol, and third-party peripheral device poses a potential security risk. Will this require an Apple-like closed environment to control the situation? If so, whose Big Brother will be granted the keys to Oz. If you recall, even Oz was hacked by a young girl with a bump on her head. Think of the business armature that currently goes into multiple auto makers and their legions of component manufacturers.
I believe that the importance of the outcome will drive our getting there in record time. But this in no way minimizes the scale of the undertaking. As they say, It’s a moon shot.
What do you think? Will we do it? What are the single greatest challenges?