Self driving cars

Self-driving cars have face two related problems; reliability and liability. First, it might make sense to distinguish between two kinds of accident - one in which the driver is injured or killed, the other in which another person outside the vehicle is. Those who get into the vehicle do so (hopefully) understanding the risk, so caveat emptor. Those who not in the car and are involved in an accident have a more substantial claim, since they did not chose to weigh the risk before hand. For these folk, the car maker's liability is likely to be much higher.

Self-driving car thus face two separate problems; one is reducing failures to the point that car buyers will assess the risk as acceptable and drive the product; and the second is reducing failures to the point that the profits are sufficient to cover potential liabilities, which means either pricing them high or reducing failure rates.

The problem is exacerbated by the fact that auto makers may be seen as having deep pockets (and thus good sources of income for lawyers). That will inhibit self-driving cars' roll-out. One solution might be to separate the car makers from the makers of self-driving control systems, which means separating the two components.

A public API that all car makers agreed to would provide an inviting platform for smaller companies to develop self-driving systems on. By having self-driving systems made by smaller start-ups, the population of  self-driving systems firms could survive the demise of one or more firms without that industry niche being crushed by litigation.