Scientific American ran a “Technofiles” piece trying to debunk the concept of auto hacking. The online version corrects errors made in their November 2015 issue where the variation of the article overstated the time required, understated the number of potentially ‘at risk’ cars, and mis-stated the proximity required to accomplish the feat.
This has been a topic here before – so I won’t repeat that perspective. However, I will copy my reply to the article posted on the Scientific American web site, since I think that this effort to dismiss the risk does a poor service to both the public, and to the industry that needs to give serious consideration for how they manage software and communications that can affect the health and safety of consumers.
Letter to editor on auto hacking
“David, et al, are not getting the message.
Yes, some of the details are wrong in David’s article (I guessed they were without being party to the Wired article) … also wrong is the “Internet” connection required assumption — external communications that can receive certain types of data is all that is required. (OnStar does not use the Internet) and the “premium savings” device advocated by my insurance company (“oh no, our folks assure us it can’t be hacked”) connects to the diagnostic port of the car (i.e. ability to control/test all aspects of operation) and is cell-phone connected to whomever can dial the number.
This is not model specific since all OnStar and after-market components span multiple models and multiple suppliers. This is not internet specific, but truly remote control would require either the cellular or internet connectivity (WiFi and Blue tooth, which are also likely “bells and whistles” are proximity limited.)
This does not require purchasing a car… they do rent cars you know. And to the best of my knowledge no automobile manufacturers have licensed software engineers reviewing and confirming a “can’t be done” — even if they did patch the flaw that the U.S. DoD/DARPA folks exploited for Sixty Minutes. — Until 9/11 no one had hijacked a commercial jet to destroy a major landmark before, so the lack of examples is not a valid argument. We have multiple proofs of concept at this point, that significantly reduces the cost and time required to duplicate this. There are substantial motives, from blackmail to terrorism (a batch of cars, any cars – terrorists don’t need to select, going off the road after a short prior notice from a terrorist organization would get the front page coverage that such folks desire.) The issues here, including additional considerations on privacy, etc. are ongoing discussions in the IEEE Society for the Social Implications of Technology … the worlds largest technical professional society (IEEE)’s forum for such considerations. see who-is-driving-your-car for related postings”
I’m not sure the editors will “get it” … but hopefully our colleagues involved in developing the cars and after-market devices can start implementing some real protections.
A question for a broader audience: “How do cell phone or internet based services (such as On-Star) affect your potential car buying?”