Can we assume the team is intelligent and has general experience with security? Can we assume that the autonomous driving Google team is aware of car-jacking? Can we assume they are able to find a better solution than human driven vehicles?<p>I say it's up to your friend to construct an argument against the SDC engineering teams being aware of the problem or unable to solve it better than the solution we already have. I have a hard time constructing either.<p>Extraordinary claims require extraordinary proof.
If there's no manual override, then it would be trivial to stop the car and threaten the passenger. According to the LA Times [0], Google is trying to make their autos fully autonomous, without any user controls. I'm not sure if there's more recent news than this.<p>If there is a manual override (wheel and pedals), the door is already locked when in motion, so the driver should have enough time to floor the gas and escape.<p>Unless Google incorporates some seriously advanced detection and response, I see SDCs being an easy target for criminals.<p>[0] <a href="http://www.latimes.com/business/autos/la-fi-google-car-20140529-story.html" rel="nofollow">http://www.latimes.com/business/autos/la-fi-google-car-20140...</a>
If someone stands in front of one of these cars then I imagine it will stay still. Someone could probably commandeer it at that point. A person in the same situation might sense they are in danger and try to get away. I would be ok with having the self driving car give itself over since that is safer for the public, but what if there is a human passenger in the self driving car?