First, I have a bias. I used to work for Boeing, but not in design or flight standards; I was a lowly computer programmer/analyst.
I fail to understand the European design philosophy in this case. When two or three pitots do not agree, why not extrapolate an air speed factor from the GPS? In other words, store the latest ground speed and also the latest difference in ground speed and air speed, both of which are (or were recently) available, and, upon pitot failure, use the last figures to generate an artificial air speed figure to allow the plane to continue on course, and not close down as if they were landed. At the same time there should be a buzzer warning to the pilots that they're not on pitots any more, but a generated air speed. This is how I would design the system.
It is solely Air France at fault with pilots unions and associations advising about the faulty Thales pitot tubes many months before the crash and Airbus advising AF of the issue over a year before hand. Air France is at fault for lack of pilot training and operating an aircraft in a known dangerous condition. There is an organisational culture within the French aviation community and the BEA protecting vested business interests and national pride over safety. Many other variables contributed to the accident but the chain of events started with complacency in pilot training and complacency in maintaining the aircraft. It is also questionable with all the technology the French have they could not locate the wreckage or black boxes and after immense pressure from the Families of the victims Woods hole institute (Basically oceanographcy organisation) finds it within two weeks without black box pingers. Irrespective of supercooling identified as a probable cause of the pitots freezing the question is would the replacements have overcome this. A personal opinion non feedback sidesticks (Joysticks) and the non moving throttle levers of the Airbus do not help an emergency situation.