Re: Pointless and expensive technology
Isn't the obvious reason for a handover a transition from an automation-friendly main road network to a minor road?
I'm quite prepared to suggest that on a motorway or dual carriageway with slip-roads, a robot can already do a better job of driving safely than an average human. Now add in the car communicating with the car in front and with the warning signs. Also add non-visual lane marking technology that works even when the road is wet and dazzly. I think that with such a near-future highway it's perfectly sensible to identify a destination, hand control to the car, and expect it to give you a minutes-long long countdown to when it will be leaving the robot-compatible highway network and requiring you to assume responsibility (potentially hours later).
Where the idea of a fully automatic car is hubristic, is on rural or urban minor roads with more unpredictable hazards. Here, a human driver and a robot working as a team would be best. Let the human do the driving, so as to remain alert. The robot may still be better at slamming on the brakes in an emergency situation such as a vehicle emerging from the side without giving way -- it's faster -- and it can alert the cars behind far faster than the human visual system can -- but it may completely fail to identify some other potential hazards as hazardous, and require a human override for non-hazards such as small fallen branches or cardboard boxes on the road (Note: I'm assuming a competent human. Many aren't!)
Once the car is on auto, it cannot be acceptable for the manufacturer to escape liability by giving a human driver mere seconds to assume responsibility. We are no good at doing that! The car must always be capable of coming to a safe automatic stop if the driver has not confirmed that he is back in charge (ideally in an escape lane to deal with human drivers who have fallen deeply asleep, without blocking a motorway exit for everyone else! )