The automotive industry is facing several changes that will shape the future of mobility and production. The car of the future will be electric, connected, and automated, and it will provide benefits for individual consumers and society as a whole. One major message of the recent Automotive in Transition Conference in Budapest was that the automation revolution is bringing challenges, but it is also bringing new opportunities for Hungary to emerge stronger from the transition process.
What will the future bring? Cars will be equipped with both internal and external sensors which will collect an immense amount of data. To be able to enjoy the full service of a connected car, we will have to disclose all this data, which raises questions about what data should be stored, where and who should store it, and how and in what context it should be used? Some original equipment managers (OEMs) use proprietary platform solutions where all data is stored in one place, and it belongs to the OEM. Alternatively, neutral platform solutions want to make data available to every interested party. The HERE network uses a hybrid-type platform, which is only open to members of a specific network.
As for the future of autonomous cars, if AI is to be used in a driving system, it has to be as good as a human driver or even surpass a human’s capabilities. How can we make sure that AI will always act in full compliance and make decisions in line with our ethical values? The EU Commission recently published a draft “Ethics Guidelines for Trustworthy AI” in support of human-centric AI. This policy paper should give AI developers a practical list of considerations in regard to important ethical decisions. But according to the Commission, ethics is not enough. The High-Level Expert Group stresses the importance of trustworthiness and technical robustness.
Another important question is: who is liable if something goes wrong? In this context, regulators and stakeholders must carefully consider and discuss the liability regime for AI systems so that society can take full advantage of AI’s innovative benefits. This applies to the use of AI in self-driving cars, but also for the use of AI – or robots – in manufacturing.
Regarding highly or fully automated cars, the legal situation seems to be clear in those countries in the EU governed by the concept of keeper liability. In general, product liability rules give the person who suffers damage a claim against the manufacturer. For cars, keeper liability grants risk coverage in addition to the product liability regime and makes the keeper liable for damages caused by the car even if he is not in fault or acted negligently. The driver of the car has no liability if the driving system is driving as long as he fulfils all obligations to take over control if the system asks him to do so. According to the amended regulations of the Vienna Convention on Road Traffic, which has been ratified by Hungary and the majority of the EU member states, there must be a driver in a car, but he can be replaced by a driving system as long as he is able to regain control. Currently, there is an amendment to the Convention being discussed which could pave the way for driverless cars on our streets in the future.
But what about AI used in robots in a plant where there is no registered keeper? Who is liable if robots cause damage? In general, our liability regime is structured in a way that any action can always be attributed to a human being. Machines have always been seen as tools of their manufacturers. However, robots might not be considered tools of the producer anymore if AI makes decision based on learned behaviour, which cannot be foreseen by the manufacturer and cannot be fully explained afterwards. Tort law rules require that there be an actor who causes damage, but who exactly is “acting” in the case of AI? The user of the robot? The owner? The producer? The programmer? The system itself? Is it justified to make a manufacturer ultimately responsible for systems based on the product liability regime? If the number of autonomous decision-making systems increase in the future and tort law rules give no valid results, strict product liability rules would apply, which would lead to a massive liability shift towards the producer that could be an obstacle to innovations entering the market. Therefore, the EU commission is currently reviewing product liability law and the liability regime. A study will be published in the middle of 2019 with proposals to create harmonized liability rules. Creating legal certainty around liability issues within the EU will also strengthen Hungary’s position as a global AI development center.
By Martin Wodraschke, Co-Head of Auto-Tech Group, CMS
This Article was originally published in Issue 6.2 of the CEE Legal Matters Magazine. If you would like to receive a hard copy of the magazine, you can subscribe here.
