Quote:
Originally Posted by Ferd
I see more issues... just one for example. the CIA was outted by Wikileaks as having a plan to hack a car to cause a crash for assignation purposes.
enter AI and the self driven car. think of the havoc some nut job will be able to cause by wrecking a dozen self driven cars during rush hour.
|
No more than the havoc the same nut job can cause ramming an old fashioned human controlled car into a group of pedestrians. Except it drastically raises the skill floor required to achieve such an attack, thus decreasing the frequency.
Right now all it takes is a person who can steer a car and hit the gas.
The scenario you proposed requires an extremely skilled hacker, potential months of preparation to get back doors in place or find some flaw to exploit, and the ability of said hacker to overcome teams of people devoting their careers to the security of the system.
There are valid reasons to be apprehensive about AI cars but safety and terrorist attacks are not among them. It would increase safety/security on those fronts across the board.
One of the biggest problems in my opinion is the ability to transition between roads where AI is expected to be in use, and back roads/off map dirt roads where the car must be driven manually. At first this would not be an issue, but as the ubiquitousness of AI cars increases you may see things like stop signs and traffic lights going the way of the dinosaur. A human having to enter this type of computer controlled traffic manually is a bit of a logistical nightmare.