Google’s self-driving vehicle crash — who’s liable when technology intervenes?

Google’s self-driving car has apparently been involved in an accident. Fortunately no humans were hurt. And fortunately for Google, a human (rather than Google’s autopilot) was driving the vehicle at the time. Reflecting on this accident in an article for Popular Mechanics, Mary Beth Griggs writes:

“[I]n ordinary, modern traffic accidents with human drivers, once fault has been established in an accident, there are two possibilities for who’s to blame — the driver or, rarely, the manufacturer, if a car defect caused the wreck.” (my emphasis)

But I think that this over-simplifies the issues involved in the current state of affairs. Today’s human-driven vehicles already often incorporate numerous technological features that (e.g.) prevent the car’s wheels from losing traction during breaking – or to regain traction if it has been lost – or to prevent the car from rolling if you’re driving fast and attempt a particularly hard turn. These are features, not defects. (See this Wikipedia entry for more information about some of these technologies.) Nevertheless, what if an accident were caused by the interaction between the human driver’s driving style and such a feature? Alternatively, suppose that the accident would not have occurred if the technology had not intervened — i.e. that they are a competent driver who knows better than the technology how to get themselves out of a pickle, and that the technology that took over was actually less capable of controlling the situation than they were. Who should be liable in this kind of situation? The driver? Or the manufacturer of this technological feature?

If we grant the premise that a human driver could be more capable than such a technology – think, for instance, of highly experienced racing car drivers who are used to controlling their vehicles in precisely these kinds of situations and who know how to regain control of their vehicles – then to me it seems that even though the accident was not the result of a defect per se (rather, it was the result of a feature), the driver should still be permitted the defense that they are not to blame. After all, if only the “feature” had not intervened then their superior skills would have saved the day (remember, we’re granting this assumption) and the car accident would not have occurred.

The problem here, it seems to me, lies in the implicit assumption that a particular technology is necessarily a feature for everyone. However, even if such a technology should be considered a feature or enhancement when judged against the (in)competence level of most drivers, I see no reason to classify it as a feature when it is compared to the competence level of highly skilled drivers.

Consequently, I see no reason to suppose that car manufacturers should only be deemed responsible for accidents caused by defects.