I challenging your statement that there would be no liability issues. I was challenging your estimation that the your anticipated savings aren't there. The facts of their QA 72 flight involve a piece of software that subjected to some of the most rigorous testing available as a life safety critical system. The liability to Airbus would have been significantly increased if the aircraft did not have two redundant safety systems in the form a pilot and co-pilot who were trained and spring-loaded to take over from the aircraft in the event of a computer failure. If the plane had been "pilotless", it is reasonable to assume the computer failure would have caused the loss of the aircraft and all passengers. The liability would have been 100x as much in that case.
Again this was a rigorously tested piece of software that had only two 3 second failures in 28 millions hours of flight time.
Now compare that testing to cars. In 2010, 20.3 million cars were recalled for safety issues. There were 648 safety recall campaigns. (http://www.dailyfinance.com/2011/01/19/auto-industry-recalls-hit-six-year-high-in-2010/)
If 1% of those safety recalls result in an injury, that is 200,000 injuries. That remains a fairly respectable accident rate.
There was also the effect of "alternate law" that the computer system switched to in light of the pilot's actions. This shifted how the plane would respond to inputs that the pilot was making. As well, Airbus aircraft average the pilot inputs, i.e. if the captain pulls up on the yoke while the co-pilot pushes down on the yoke, the system will average the inputs and fly level (an oversimplification admittedly but I'm using it just to show the human factors problems). Other manufacturers of aircraft elect to have differing inputs provide feedback -- i.e. if the captain it pulling back and the co-pilot is pushing down, they feel the opposite inputs in the control yoke.) The other crew didn't realize until it was too late that co-pilot was in a climb stalling the aircraft because the software didn't provide that feedback. Clearly, if that same inexperienced crew were flying a different aircraft, the crash **might** have been avoided because the software would've have handled the inputs differently. So how much did poor software design contribute to the accident?
My response is that aircraft companies get sued for failures of the their automated systems. The Quantas Air Flight 72 is a good example of this. Millions were paid out in damages from Airbus and Northrup Grumman. Keep in mind that the system that failed "had only failed twice in 3 seconds despite 128 million hours of successful operation" according to one aviation source. As well, passengers were partially responsible for their injuries because they were advised to keep their seat belts fastened. None of that stopped a successful law suit in Australia follow by a second lawsuit in the US for passengers that wanted a larger payout.
Extrapolate that to volume of potential of "3 second failures" in driverless autos and that's a huge liability.
Many things today are quite safe, yet the manufacturer's of those things get sued. Late-night TV is filled with ads for law firms looking for patients harmed by medical procedures or medications. Medical procedures carry inherent risks that the recipient acknowledges in their consent for a procedures. Medication inserts document that x% of the test population experienced negative side effects from taking this medication. Yet successful lawsuits happen regularly. The reason is the lawyers don't argue whether the procedure was risky but whether the pharma companies or surgeon did everything possible to mitigate and minimize that risk.
While I understand the "computers would perform better than people" from a practical point of view, the standard that will be used in the court of law is whether the programmer exercised sufficient due diligence to insure the mistake the computer made could have been avoided. You can testify on the stand that the car may perform better than a human in a similar case. The lawyer will respond that we are not talking about how the car performed but how the programmer performed. Would a different programmer have written software that would have avoided the accident? If so, why didn't you hire that programmer and spend the money to avoid my client's injuries. You are just an evil penny-pinching corporation recklessly disregarding my client's safety! Your computer software is the modern equivalent of the Ford Pinto!
That the lawsuit that car manufacturers fear because that is the one that they will face in court.
In fact that is the case in NJ. Corporations are using the term "civil union" as a defense against providing the rights under the law. The argument goes "if the government intended you to have marriage rights they would have granted marital status." In the case of UPS, it required the governor to apply political pressure for them to abandon that argument.
It is pride is success of surviving despite some of the difficult challenges we faced growing up and still do today. I'm proud that despite that growing up hearing faggot tossed around every day in school, that I managed steel myself against the taunts and make it through school and go to college. I'm proud that despite facing groups of guys driving on the street yelling slurs and threats that I'm not afraid to leave my house. I'm proud that I was willing to be out despite the potential loss of my career. I'm proud that I was able to drum up the courage to come out to family, not being totally sure that I would be disowned. I'm proud of the fact that I no longer let the sideways glances and disapprovingly looks bother me any time I'm be out with a boyfriend.
Finally, I'm proud that I have the courage to stand up and fight for the rights of my brothers and sisters so they don't have to face what I faced.
Slashdot Mirror
Again this was a rigorously tested piece of software that had only two 3 second failures in 28 millions hours of flight time.
Now compare that testing to cars. In 2010, 20.3 million cars were recalled for safety issues. There were 648 safety recall campaigns. (http://www.dailyfinance.com/2011/01/19/auto-industry-recalls-hit-six-year-high-in-2010/)
If 1% of those safety recalls result in an injury, that is 200,000 injuries. That remains a fairly respectable accident rate.
There was also the effect of "alternate law" that the computer system switched to in light of the pilot's actions. This shifted how the plane would respond to inputs that the pilot was making. As well, Airbus aircraft average the pilot inputs, i.e. if the captain pulls up on the yoke while the co-pilot pushes down on the yoke, the system will average the inputs and fly level (an oversimplification admittedly but I'm using it just to show the human factors problems). Other manufacturers of aircraft elect to have differing inputs provide feedback -- i.e. if the captain it pulling back and the co-pilot is pushing down, they feel the opposite inputs in the control yoke.) The other crew didn't realize until it was too late that co-pilot was in a climb stalling the aircraft because the software didn't provide that feedback. Clearly, if that same inexperienced crew were flying a different aircraft, the crash **might** have been avoided because the software would've have handled the inputs differently. So how much did poor software design contribute to the accident?
My response is that aircraft companies get sued for failures of the their automated systems. The Quantas Air Flight 72 is a good example of this. Millions were paid out in damages from Airbus and Northrup Grumman. Keep in mind that the system that failed "had only failed twice in 3 seconds despite 128 million hours of successful operation" according to one aviation source. As well, passengers were partially responsible for their injuries because they were advised to keep their seat belts fastened. None of that stopped a successful law suit in Australia follow by a second lawsuit in the US for passengers that wanted a larger payout. Extrapolate that to volume of potential of "3 second failures" in driverless autos and that's a huge liability.
Many things today are quite safe, yet the manufacturer's of those things get sued. Late-night TV is filled with ads for law firms looking for patients harmed by medical procedures or medications. Medical procedures carry inherent risks that the recipient acknowledges in their consent for a procedures. Medication inserts document that x% of the test population experienced negative side effects from taking this medication. Yet successful lawsuits happen regularly. The reason is the lawyers don't argue whether the procedure was risky but whether the pharma companies or surgeon did everything possible to mitigate and minimize that risk. While I understand the "computers would perform better than people" from a practical point of view, the standard that will be used in the court of law is whether the programmer exercised sufficient due diligence to insure the mistake the computer made could have been avoided. You can testify on the stand that the car may perform better than a human in a similar case. The lawyer will respond that we are not talking about how the car performed but how the programmer performed. Would a different programmer have written software that would have avoided the accident? If so, why didn't you hire that programmer and spend the money to avoid my client's injuries. You are just an evil penny-pinching corporation recklessly disregarding my client's safety! Your computer software is the modern equivalent of the Ford Pinto! That the lawsuit that car manufacturers fear because that is the one that they will face in court.
In fact that is the case in NJ. Corporations are using the term "civil union" as a defense against providing the rights under the law. The argument goes "if the government intended you to have marriage rights they would have granted marital status." In the case of UPS, it required the governor to apply political pressure for them to abandon that argument.
It is pride is success of surviving despite some of the difficult challenges we faced growing up and still do today. I'm proud that despite that growing up hearing faggot tossed around every day in school, that I managed steel myself against the taunts and make it through school and go to college. I'm proud that despite facing groups of guys driving on the street yelling slurs and threats that I'm not afraid to leave my house. I'm proud that I was willing to be out despite the potential loss of my career. I'm proud that I was able to drum up the courage to come out to family, not being totally sure that I would be disowned. I'm proud of the fact that I no longer let the sideways glances and disapprovingly looks bother me any time I'm be out with a boyfriend. Finally, I'm proud that I have the courage to stand up and fight for the rights of my brothers and sisters so they don't have to face what I faced.