Wednesday, 5 November 2014

Don't Be Shallow Hal About Your Driving Decisions--I Mean, Maybe An Autonomous Car Isn't Such a Good Thing After All...



By 2020, the option of having a self-driving vehicle will be a reality--that's only six years away. According to According to Kurzweilai.net, most auto manufacturers are "incrementalists," adding automated features such as adaptive cruise control, self-parking, and traffic-jam assist. But, what we know about technology is that it is spiky, not linear. this means that the latest technology builds on the advancements and strength of the previous technology. It thus can spread very fast.



Ray Kurzweil on the exponential growth of computing

There are three options when looking at automated vehicles:

1. Partially automated: Here a car can take over some aspect of driving, as we've show in other posts, such as speed and steering, but cannot control much beyond that;

2. Highly automated: The car can drive itself, but is also equipped with a human over-ride system;

3. Fully automated: The car drives itself without human over-ride.

Don Norman, a professor of design from University of California, San Diego and author of the Design of Everyday Things, gave a lecture endorsing Google's project for fully automated vehicles in which he claimed that a highly automated car would be less safe than a fully-automated one. Airplanes, Norman pointed out, have been highly automated for years; however, the over-ride system provides several minutes for the pilot to take over control. The problem with cars, at least thus far, is that one would have only a split second or two to take over, should the vehicle be unable to handle driving conditions.



Don Norman, The Design of Everyday Things

Has the Google forecasts been pre-empted by Tesla, whose new vehicle, as we've seen in a previous post, the Model S already has an autopilot setting that was brought to market faster than anticipated. It consists of four parts:

1. Long-range radar that can see through anything,

2. Camera with image-recognition, so that it can read stop signs, distinguish pedestrians,

3. 360 degree long-range ultrasonic sonar (how cool does that sound!) that establishes "a protective cocoon around the vehicle and make a smart move by looking at the ultrasonics, and

4. The navigation and GPS and real-time traffic.



Elon Musk and Tesla Model S


All this sound way too cool, especially for a quasi-Ludite like me. But the question that some are asking is whether this move toward autonomous vehicles is actually something that we as free human beings want. It sounds cool, it sounds convenient (I mean, why wouldn't you want your car pulled out of the garage 2 minutes before you have to leave for work, with the climate set at a perfect 73 degrees and your favourite Duran Duran tunes running in the background?), but are we getting ourselves into a situation in which our freedom to drive is being slowly taken away? 

If a computer can drive safer than you, and put less people at risk on the roads, then as a human you are a liability, and thus unfit to drive. In 2020 are we going to be facing these legal issues? Will insurance companies lean toward insuring 'Hal' the vehicle automated driving system over you the human being? And if so, where is your freedom to drive? Is driving even something worth having freedom to do? 

These are very important questions to ask ourselves, again, as we are dealing with exponentially growing, unpredictable technological tools.


No comments:

Post a Comment