REASONS TO BE SKEPTICAL ABOUT THE FUTURE OF SELF DRIVING CARS:
1. The issue of greater safety has not been demonstrated, nor can it be until the cars are on the road. One billion miles in a demonstration? That’s just a good start. Americans drive billions of miles per year and a full, reliable test period would probably last at least three years with tens of thousands of cars. Why else should rely on these unverified claims? We shouldn’t.
2. We have no idea how cars with drivers and cars without drivers will interact. Will the drivers learn to cut off or otherwise frustrate computer driven cars?
3. If people are being driven electronically most of the time, their driving skills will go down dramatically. That means when they have to take over, there will be increased danger, not less. Driving skills and improved cars are the main reasons that we don’t have ten, twenty or thirty thousand more people killed on the roads every year.
4. You can NEVER speed. The car in concert with programmed operating instructions will decide how fast you can go. Unless driverless cars mean that speed limits can be increased, the great American sport of speeding will be over and people will be greatly frustrated. 80 to 90% of drivers speed now a majority of the time. All you have to do is pay attention and you see it everyday.
5. Driverless cars will shutdown on their own and it will be very difficult, if not impossible, for someone who isn’t a technician to get them going again. Early personal computers were comparatively simple things and a dedicated user could often find a work around when the computer refused to do what you wanted. As the computers have gotten more common and complicated, the ability of users to understand the system and to re-direct it have become much more difficult. PCs are increasingly built to instruct you what to do, not allow you to use it in the way you want. This will happen with the cars, too, making them nearly impossible for the average person to re-direct or trick into doing what you want, like starting and taking you to work if one system is failing.
6. A driverless car would be highly vulnerable to robbery. One guy stands in front, another stands in back while a third holds you up, What would you do? How could you escape?
7. What about hydroplaning? One rule in driving now is never use cruise control in heavy rains because if the wheels start slipping, the accelerator stays down, most likely spinning the car out of control and into a major crash. Getting out of hydroplaning requires great skill and subtlety. First, let off the accelerator. Second, turn the steering wheel, very slightly, in the opposite direction from which the car is starting to spin. Third, wait for the hydroplaning to stop and begin driving normally again. How could a computer handle those moves? If it goes wrong on any part, people will die.
8. If safety hasn’t been demonstrated and, potentially, doesn’t work out, what’s the point? Presumably, there would be far fewer crashes related to alcohol and people falling asleep at the wheel. Most deaths occur late at night on two lane roads, so, if driverless cars were safer there, that would be a big gain. What if they weren’t?
What if driverless cars encouraged people to get drunker and get in a car? It is likely that more trouble would be caused elsewhere by such drunks. The normal limits on human behavior help to form barriers around excessive actions. With driverless cars, a huge barrier would be removed.
9. Where is the cost/benefit analysis? Every choice involving dangers and transportation involves trade offs, decisions about how many lives can be save at what cost. What would be the total added cost of making the entire number of American cars work without drivers? The companies investing in this technology don’t care, because they think most of that money will be coming to them.
10. Would electric cars work less well with all that computer technology also sucking electric impulses? If so, where’s the trade off there and is it worth it?
Doug Terry, 10.15.17