Photo: YouTube screenshot
The idea sounds completely ridiculous coming from me, but what if I told you it actually belongs to someone who writes for MIT's paper? Hmm, maybe you were a little too hasty there to dismiss it after all.
Right now, the manufacturers, the press, the regulators - the everybody makes it sound as though the thing standing in the way of autonomous cars is their ability to navigate the roads safely. In other words, when they'll be able to get from A to B without making a street post look like a C in the process, they're good to go.
Simson Garfinkle (yeah, I know, cute parents) of MIT Technology Review would beg to disagree. He believes that, like in so many other cases, the developers of this technology have gone about it a bit backwards.
He argues that security against all types of attacks that could be directed at the AI-controlled vehicles should have been the primary concern, and not something left to figure out after the technology is already out on the streets.
Past experiences don't cast a very favorable (or comforting) light on the future. Mr. Garfinkle brings up the beginnings of the electronic mail system and the spam it brought with itself. It took us over a decade to contain the problem, and even today you are just one missed click away from finding how to enlarge your penis instead of opening a message from your mother.
People might not realize just how sophisticated autonomous cars are, the immense computing power they require and the very advanced artificial intelligence they rely on. Blocking all the access points for potential hackers becomes increasingly difficult with the growing complexity of a system, so on that basis, autonomous cars make for especially inviting victims.
And if hacking was largely restricted to stationary objects that usually sit on desks so far, now we're talking about heavy machinery moving at considerable speeds. The effects of any malicious interference with the AI has the very high potential of resulting in the loss of human lives. Which, ironically enough, is exactly why we were making self-driving cars in the first place.
As for who would want to cause such issues, Mr. Garfinkle makes a very straightforward and human observation: the people whose jobs the robotic cars would take over. And even though we said something about becoming hackers, if things don't change over the next years, the truth is that all these angry unemployed people might need to do is flash a laser pointer at the vehicles' cameras or throw some bird poo on a stop sign.
Simson Garfinkle (yeah, I know, cute parents) of MIT Technology Review would beg to disagree. He believes that, like in so many other cases, the developers of this technology have gone about it a bit backwards.
He argues that security against all types of attacks that could be directed at the AI-controlled vehicles should have been the primary concern, and not something left to figure out after the technology is already out on the streets.
Past experiences don't cast a very favorable (or comforting) light on the future. Mr. Garfinkle brings up the beginnings of the electronic mail system and the spam it brought with itself. It took us over a decade to contain the problem, and even today you are just one missed click away from finding how to enlarge your penis instead of opening a message from your mother.
People might not realize just how sophisticated autonomous cars are, the immense computing power they require and the very advanced artificial intelligence they rely on. Blocking all the access points for potential hackers becomes increasingly difficult with the growing complexity of a system, so on that basis, autonomous cars make for especially inviting victims.
And if hacking was largely restricted to stationary objects that usually sit on desks so far, now we're talking about heavy machinery moving at considerable speeds. The effects of any malicious interference with the AI has the very high potential of resulting in the loss of human lives. Which, ironically enough, is exactly why we were making self-driving cars in the first place.
As for who would want to cause such issues, Mr. Garfinkle makes a very straightforward and human observation: the people whose jobs the robotic cars would take over. And even though we said something about becoming hackers, if things don't change over the next years, the truth is that all these angry unemployed people might need to do is flash a laser pointer at the vehicles' cameras or throw some bird poo on a stop sign.
Google News