Artificial intelligence is invading our lives, not to mention controlling them. Look around and you’ll find a connected object, an element devoid of a simple form of intelligence, capable of responding to you when you speak to it, or simply adapting to its environment, like robotic vacuum cleaners . For those with a fairly recent vehicle, you even have a host of driving aids (most of which have become mandatory) that lay the foundation for tomorrow’s autonomous driving. But while the first tests on open roads begin in Germany and France, the question of liability still arises in the event of an accident. And until today, no one has been able to answer that.
Autonomous car accident, who is responsible?
And that’s the million dollar question that no one can really answer. It is clear that the manufacturers who put “autonomous” cars (not fully, but level 3) on the road today must have communicated on the subject. One thinks in particular of Mercedes, which has circulated for some time the S-Class and EQS equipped with the new Drive Pilot technology. The German brand clarified things from the start: in the event of an accident, it is the manufacturer who will take responsibility if the mode was activated under the appropriate conditions.
Yes, but here, behind this simplistic facade, hides a much more complex reality.. Manufacturers never work alone on such projects, which bring together the brand’s experts and engineers, but also equipment from suppliers and sometimes even lines of code written by developers working for third parties. So in the event of an error or accident, who pays the bill? The builder alone? The manufacturer and equipment supplier involved in the failure? This is exactly what the EU Commission is trying to clarify.
Make life easier for complainants
A European directive protects people in the event of material damage caused by a defect in the product. They can then hold the manufacturer responsible. But it dates back to 1985 and is no longer suitable for the world we live in, which is increasingly mixing in AI.
This is well summarized by a European lobby group: “As we see with AI, new technologies create interdependencies between developers, suppliers and users, which can blur the lines of who is responsible for what. So for AI, we need to bridge these legal gaps to better be able assess the individual contribution of each actor and an appropriate line of causality“.
In order to better break down the various scenarios, several scenarios arise according to the EU :”These are cases where the manufacturer fails to provide the information elements or if the product does not meet the safety requirements. It is also the case of obvious malfunctions or if the causal relationship is impossible to prove due to the technical or scientific complexity of the product. This last scenario aims to prevent the “black box” effect of AI systems that may be beyond the understanding of their own developers. In these cases, the complainant only needs to prove that the AI in question contributed to the injury and that the product is likely to be defective.“. Each time you will notice that the person responsible is never the driver. But only if he has used the autonomous driving technology under the right conditions! It is clear that drivers who have fun, especially overseas Atlantic, to get in the back seat of their Tesla while the car is driving alone would theoretically not be covered and therefore 100% liable.
Finally, these directives adopted by the EU aim to make life easier for people who experience incidents with an autonomous vehicle or any object equipped with an AI. Complainants will not have to “describe and demonstrate” the malfunction in the AI. They may even have access to certain internal manufacturer data to confirm the brand’s level of responsibility. Suffice to say, automakers will need to pay attention to the development of their autonomous driving systems.