Attacking AI based systems with malicious input is like shooting fish in a barrel. Combine high complexity with low comprehension of how the internals of the system actually work and you have a field day for security researchers.
It’s only a matter of time before someone hurts or robs a Tesla driver by forcing the car to do something that’s not in the best interest of the operator.
If self driving cars become disabled by putting a cone on the hood, maybe they aren’t ready to be on the streets.
They won't be ready to be on the streets until they've had a lot of time on the streets.
It's a catch 22.
I wouldn’t expect a human driver to move if their view is obstructed or there are objects on the vehicle
Humans are capable of assessing and addressing the obstruction; meanwhile these cars are permanently disabled without outside assistance.
Attacking AI based systems with malicious input is like shooting fish in a barrel. Combine high complexity with low comprehension of how the internals of the system actually work and you have a field day for security researchers.
It’s only a matter of time before someone hurts or robs a Tesla driver by forcing the car to do something that’s not in the best interest of the operator.
Source for the article