Dear Colleague,

Last week I read an article published on 12 July 2016 online by the Fortune magazine: "Another Tesla Motors Car Crashed while in Autopilot". 

This fact made me think of whether automated cars may or may not be reliable. As a matter of fact, one Tesla car was already involved in a fatal accident early May 2016 and in that case the company stated that one accident after running 200 million kms of tests could be acceptable.

The first things to be noticed is that Tesla cars are not automated cars "per se" but they become automated after downloading a piece of software. Maybe the software needs improvements?

I don't know but it can be that a vehicle designed as automated/driver-less from the beginning - as the Google car - could be different from one that becomes automated after installing a software patch.

In the latest accident, the user stated that the autopilot failed to detect debris in the road while the car was travelling at 60 miles per hour (96 km/h). Tesla denied that the crash was caused by an autopilot failure but instead it was due by an incorrect driver's behaviour (more details in the article above).

Whatever real cause was (maybe it will come out more clearly in the future), the optimistic vision of automated, driver-less vehicles that can circulate safe in cities (someone speculated traffic lights would not be necessary anymore as automated cars will self-regulate the traffic) must be re-assessed and enthusiasms calmed down. Additional analyses, studies and tests certainly need being carried out. 

Insurance companies will also have to take into account whether a vehicle was originally designed as driver-less or it was upgraded so: such a nightmare...