[thedoc (OP)]

Jeff Kirtland asked the Naked Scientists:
   
The interview regarding liability and driverless cars missed a critical point.  Currently, the blame for accidents is recorded as human error, I would argue, largely because the human is legally and morally responsible.  In fact, we know that driverless cars are currently killing people at a higher rate than human driver cars (deaths in the US occur at roughly 1 per 100 million miles driven—driverless cars have only covered a small fraction of that).  What is worse, is numerous people were killed in Toyotas transposed the accelerator with the break.  Because the computer recorded the action of the peddle not the intended function of peddle the deaths were attributed to human error when it was a computer error.  I have an example that may help.  The insert key on my key board is mapped as the delete key.  Any key stroke monitoring software would record delete every time I hit insert.  This was the same with the Toyota deaths.
 
My point, we assume that accidents on the road are human caused because that is what we record.  Furthermore, we are assuming that if all cars followed all the road rules—including the contradictory rules—accidents would be eliminated.
 
It  would be nice if 5 Live could look into whether these assumptions are in fact robust enough to be staking our lives and motoring future on.
 
Remember, I am waiting to have my Tocata airbag replaced three years after the problem was identified.
 
Regards,
 
Jeff Kirtland
What do you think?

[Marika]

With the increasing popularity of driver-less cars, and an accompanying increase in driver-less car accidents, there have been more and more legal cases looking into this issue. The current stance is that driver-less car technology exists to assist the driver, not to replace them. So the driver is still responsible to pay attention to the road and overtake the auto-pilot feature in the case of an emergency.