In an attempt to short-circuit what might be protracted considerations by governments on the legal and liability issues relating to the use of autonomous cars, Volvo recently announced that it would accept full liability if one of its autonomous cars was involved in an accident due to a ‘flaw’ in the car’s design. Other manufacturers (Google and Mercedes-Benz ) have made similar statements.
The implications of this simple pledge are quite far reaching. To truly do this, Volvo would have to have full control over all aspects of the car operation, servicing, and maintenance. In fact the user would have to be in the position of having absolutely no responsibility for the car whatsoever. Almost inevitably this would imply a complete shift over to leasing rather than owning the vehicle. In fact it could well mean that any private ownership of such a car was no longer an option.
Volvo currently see their cars as being conditionally autonomous, in the sense that they would be autonomous only in certain places or under certain conditions, and to still be owned privately. In this case the human driver by default will always have some arguable degree of liability. The transition between human and software control (the ‘handoff’ problem) is a major challenge for any automated system. Be under no illusion, if an accident occurs due to a hand over problem it will not be the design of the automated system that bears the blame, but the perceived failure of the human to respond in the manner required by the system design.
So what does this statement by Volvo really mean? It is not necessarily a simple matter to determine who or what is at fault in an accident, as many accidents have a number of contributory factors. Any car manufacturer would do all it can to protect its reputation, even more so for an emerging technology.
The liability for most accidents is currently determined by the insurance companies, based on evidence presented to them by the parties involved, and in many cases will result in some agreement between the insurers to share costs, particularly where it is difficult to determine liability. The consequence of the decision is borne by the insured driver through increased premiums.
In a fatal accident it may well be the police who determine liability, though in some cases this could be decided by a court. The parties involved in making the decision of liability are independent and have no ulterior motive for arriving at that decision. An insurance company can take a statistical view of the risk of paying out, and has no axe to grind regarding the make of vehicle.
How different this would be if a manufacturer were involved, even via a third party, for the simple reason that their reputation would be at stake. An insurance company (or the police if a death is involved) may well inspect a car that has been involved in an accident, but they will only be looking to establish in broad terms whether the vehicle was in good working order. A manufacturer with their business on the line might well take a far more aggressive approach when it comes to arguing about apportioning blame.
It would be in the interest of the manufacturer to reduce to the absolute minimum the number of times that their vehicle was deemed to be responsible and to use all their resources to do so. If an autonomous vehicle, or one operating in autonomous mode, is found to be responsible for an accident it is the manufacturer who bears the blame and there are implications for their entire business. A legal contest with a global car manufacturer could be a rather one sided affair.
What if the washer fluid to keep the sensors clean was not made up according to the manufacturers specification , or maintenance or repair not carried out by an approved garage, or a routine service not carried out in strict accordance with the manufacture’s specifications and service intervals? What if the control software had not been updated?
What if the company that made the car went out of business?
Comments