Automated driving and the liability of the manufacturer

Guidelines for automated driving: consequences for manufacturers?

On 20 June 2017, the Ethics Committee, led by former Supreme Federal Constitutional Court judge Udo Di Fabio, submitted its report on automated driving. During the preceding nine months, the Committee had developed some initial guidelines for automated driving. Central to the report are 20 propositions, which deal with issues including data sovereignty, unavoidable situations and fundamental goals for the domain of automated driving. Right at the beginning, in the first proposition, the central idea that automated driving should follow is made clear: “Partly and fully automated traffic systems serve primarily to improve the safety of all those participating in road traffic.“ 

This is followed by the second proposition, which goes on to shape that idea in such a way that the objective should be a reduction in the amount of injury and damage, the ultimate aim being the complete avoidance of accidents. This is made more concrete in the second sentence of the second proposition, which states that the licensing of automated systems is only tenable “if in comparison with human driving performance it promises, at the least, a reduction in the amount of injury and damage in the sense of a positive risk balance”. The Ethics Committee has thus made it unequivocally clear what demands should be made on manufacturers if they want to license vehicles with automatic or autonomous driving systems: the automated systems must surpass the average driver in respect of driving safety and accident avoidance. It is true that this expectation is a logical extension of orientation toward a benefit in terms of safety, but it does entail a central problem.

There is on the one hand some emphasis to the effect that the automated and connected technology must be designed in such a way that threatening situations cannot develop in the first place; explicit reference is made to the dilemma situation, in which the automated vehicle has to decide which of two ‘non-tradable evils’ (losses of human life) it is to choose; whereby in other trade-off situations, according to the report, the safeguarding of human life always takes priority over damage to property and injury to animals. On the other, the report fails to offer a solution for those very cases in which the autonomous or automatic system really is confronted with a situation of that kind. The authors content themselves with pointing out that in a situation in which inestimable objects of legal protection, i.e. lives, are in mutual opposition, the decision ‘life vs. life’ cannot be programmed in such a way as to be free of doubt, and that the decision therefore continues to be incumbent on the ‘morally discerning’ and responsible driver. There is no general impunity as regards action taken in dilemma situations; the culpability of the actions of the individual involved must be examined from case to case. Yet automated action is not covered by any such exemption. So the implementation of autonomous vehicle systems is likely to depend on whether or not the system succeeds in almost certainly precluding situations of that kind.

We are, furthermore, reminded in proposition eleven that the basic principles of product liability apply ‘normally’ to automatic driving systems. This hint to the manufacturers carries the footnote that they are under obligation to optimise their products and observe the market on a continuous basis. With regard to self-determination in respect of the data generated, proposition fifteen explains that the registered vehicle keeper or the vehicle user is to decide about whether or not his or her vehicle data may be passed on and if so how they are to be used. However, this free will aspect in respect of the passing on of data should be practicable for the user of the data too. 

In conclusion, it should be noted that whilst the report of the Ethics Committee does not develop a directly binding effect, it could, indirectly, constitute an initial indicator for the legislators. That said, it does not offer an approach to solving the problem of the dilemma situation.

[July 13th, 2017]