Today we saw Mobileye take the extraordinary step of distancing itself from Tesla after the death in Florida of a Telsa driver alleged to be on auto-pilot at the time of the crash. At the same time we hear of a Chinese family preparing to sue Tesla over a second death.  

Mobileye the supplier of Telsa's currently deployed driver assistance technology issued the bruising statement that it did not want to be associated with "pushing the envelope in terms of safety".  This all hot on the heels of the July signing of Mobileye's new customer BMW and a new VW deal also rumoured to be imminent.

Tesla on the other hand said all this has more to do with the fact it intends to replace Mobileye's EyeQ3 chip with Telsa's own computer vision technology.  Tesla also says it has continuously educated customers on the use of Autopilot features, reminding them that they’re responsible to keep their hands on the wheel and remain alert and present when using Autopilot.  

One difference between a death on autopilot and death in a conventional car is liability. With 94% of conventional road deaths down to human error, there was in most cases relatively little chance of reputational risk or liability coming back to the automaker. On the other hand a death on autopilot (whether in a semi-autonomous or fully autonomous car) can more easily be blamed on the automaker or a supplier of a key component. 

Dan Tench says: “In the UK, under the Consumer Protection Act 1987, manufacturers are potentially liable for damages (including damages for personal injury) arising from any “defect” in their products.  This includes component manufacturers whose products are then included in other manufacturers’ products.  A defect will be held to exist where the “safety of the product is not such as persons generally are entitled to expect”.

"If an incident of this type were occur in the UK, the car manufacturer would potentially face a claim on the basis that the car itself had a defect.  More difficult is the position of the manufacturer of the driver assistance system.  It may say that there was no defect in the system for the purpose for which it was intended, and it was only because it was used for another purpose that there was a problem.  However, if that manufacturer knowingly acquiesced to the use of the system in the car in a way which meant that it was no longer safe, it is not impossible that it could face a potential claim under UK law.”

The stakes are going to be higher for the automakers and their suppliers with this redistribution of risk.  Despite the claims it is of course possible in a semi-autonomous scenario for a death still to be down to human error (or incorrect use of the driver assistance system). In the case of fully autonomous, there is less room for human error.  We have not yet worked out whether the insurers will see any real difference in paying out for a situation where the automaker or its supplier is at fault as compared to when a human is at fault. In principle at least the pay outs ought to be less often with machines in control.