[T]he advent of smart self-driving cars has taken over the globe by storm. With each passing day, we usually witness an advertisement or an article that is related to the innovation brought by a leading auto company manufacturer that involves getting a car to navigate autonomously. Usually, the idea gets people interested and impressed. However a short commercial released by Lexus recently that demonstrates the capabilities of a non existent feature called ‘Lane Valet’ that allows a driver to move lane hogs from hogging the fast lane at the touch of a button brings about a few thoughts into mind. Where does the idea of self driving autonomous cars conflict with the realm of privacy?
The feature was shown to work by connecting with the car in front without any consent from the driver of the vehicle and controlling it so that it would switch lanes.
Besides such a feature being scary and immensely obstructive to privacy (imagine your car being controlled by someone else simply at the touch of a button) the general enthusiasm regarding the idea of self-driving autonomous vehicles that are net connected has been positive from the community. The feature, even though revealed to be an April Fool’s prank by Lexus doesn’t seem too unrealistic in a world where the Internet of Things ensures that everything can be connected to and controlled.
Similarly, Tesla’s Autopilot also ensues the idea that driving data from your vehicle is somehow being synchronized with a cloud at certain intervals of time; since it isn’t possible to have a 24/7 data connection at all times – might be possible if the car maintains a satellite link though, the issue of privacy once again smokes up the windshield of opinion regarding the adoption of smart self-driving autonomous cars.
The idea starts with understanding how smart autonomous net connected AI-powered vehicles work. Since the amount of processing required on a smart car is in the magnitude far greater than what an on board processor would or could carry, some of the computational effort is offloaded onto a cloud server. This means that whatever data that you are sending out when your car is driving is being recorded elsewhere. Your driving patterns, your habits and the daily route you take could always be sold to the highest bidder if this data is unsecured or a nosy hacker gets hold of it.
Even if the car’s on board processor can somehow manage to sift through the millions of calculations required for effective decision making in a self-driving scenario, data would be routinely sent to a cloud connection for the purpose of improving quality and tracking defects and bugs in the AI software running the show. Most of the cars today also have FoTA or ‘firmware over the air’ update capabilities so hence it is mandatory that the AI software piloting the vehicle will definitely be updated frequently with some back end hidden or in view connectivity suite.
Companies that are manufacturing these smart cars and that haven’t as of yet ventured out into the public arena need to work on some concrete security algorithms and have established business logic that enable the securing of data that is related to the driving patterns of people that will be using this car.
The way self-driving car technology can be misused in the future is concerning. This does not mean that it become a select target for attack as any and all technology can and is being misused currently. The recent news covering how ISPs have been authorized to sell a person’s internet browsing history to advertising agencies shows when the legal cover afforded by national governments to consumers can be blown off at any time. Imagine that you’re driving data is legalized to be sold off in this manner in the future, if currently you believe that the government in your country will offer you legal protection from such an event from taking place.
Let us also for the sake of discussion invoke an alternate scenario. In this day and age, technology is rapidly attempting to reach the areas where there is little to none of it in an effort by large technology corporations to increase their number of customers. Call it a clever business strategy if you will but it works. Assuming this pace keeps up as it is, the self-driving car technology reaches every inch of the globe and becomes a norm around the globe – the authoritarian regimes in the world will have a field day. Imagine locking people in their cars or uploading everyday routes that will ensure that you go to only where the regime wants you to. This will effectively terminate dissent to the point that regimes would be able to take anyone hostage anywhere.
Something as simple as remotely locking the windows of your car would become possible by any criminal with a bad intent that manages to invade the car’s system or even the authorities of your country that wish to do so. The scenario was recently demonstrated with Tesla’s auto pilot system in which an external invader hacked into the car and applied the brakes without the consent of the user. With Facebook and Google complying with so many requests put forth by government agencies to provide data regarding its users it could very well be the case in the next coming decade with smart car providers. What is needed to be understood here is that the smart artificially intelligent car is more of a service than a product and requires constant improvement just as any software does and without sharing statistical data regarding performance at minimum and computational offloading at a maximum the service become difficult to render.
Consumers of today that will become the early adopters of this technology should take necessary measures and communicate these concerns with auto manufacturers that are leading the role in developing autonomous AI powered smart cars so that timely action can be taken with regards to the development of effective security features that protects the user from privacy invasion and misuse. Google – a leader in the area is well versed with this concern and can help charter an initial level of agreement between leading auto manufacturers that sets the ethical standards that this technology needs to be developed around.