The development of self driving cars has resulted in some very valuable progress in robotic understanding of context (your car’s surroundings while driving), but the organic nature of randomness in our world makes completely autonomous systems too risky to permit. The value of such systems as a supplementary safety tool should not be underestimated, but until they can anticipate a child chasing a baseball running into traffic from a sub grade side of a street (from between two parked SUVs) while the autonomous vehicle is being followed by an overloaded panel truck without causing an accident that an alert human might have anticipated they will never be safe enough.
Can autonomous vehicle technology help make us safer? Absolutely, in fact we should continue to support human drivers by giving them the very best of this tech as backstop against human error, but as a computer scientist I can not imagine risking any child’s life to demonstrate the limits of it’s abilities. Wishful thinking is great for driving innovation, but no human life is worth allowing you to abrogate responsibility for paying attention on your daily commute.
Questioning Assumptions 