[co-author: Paschalis Lois]
When a parliamentary report cites Mary Shelley’s Frankenstein in its recitals and proposes new regulation for robots with artificial intelligence (“AI”), one cannot be sure whether the 19th or the 21st century has inspired the legislator. Last week, the European Parliament took a step to introduce new regulation of robots in Europe. Declaring that the EU needs to “take the lead” in this area, the Parliament endorsed a Report that asks the European Commission to propose rules on robotics and artificial intelligence, in order to fully exploit their economic potential and to guarantee a standard level of safety and security. The Report addresses various kinds of robots, amongst others autonomous vehicles, care robots, medical robots, or drones.
This follows an earlier report published in May 2016 on Civil Law Rules on Robotics with proposals on the regulation of the robotics industry at an EU level. Such regulation was argued necessary in order to ensure that the EU and its Member States maintained control over the regulatory standards at which the industry operated in the EU, as well as to ensure certainty for enterprises planning to develop their businesses therein. The parliamentary Committee on Legal Affairs released a preliminary study on the impact of robotics on civil law.
Liability rules
The issue with the question of liability for robots is that the more autonomous robots are, the less they can be considered simple tools in the hands of their manufacturer, programmer, owner or user. This makes the general rules on civil liability potentially insufficient and, in the European Parliament’s view, calls for new rules which focus on who should be held – partly or entirely – responsible for the acts or omissions of a robot.
But how to determine who should be responsible for damage caused by such a machine? The Parliament suggests that an obligatory insurance scheme as for cars could address the complexity of allocating responsibility for damage caused by increasingly autonomous robots. This could allow the manufacturer, the programmer, the owner or the user of a robot to benefit from limited liability insofar as the use of smart autonomous robots would be covered by a supplementary fund to ensure that any victims do not remain uncompensated for. Moreover, the Parliament urged the Commission to ensure that a specific legal status is established for robots in the long-run so as to establish liability in the event they cause damage.
From this, one can already appreciate the legal as well as philosophical complexities that arise from regulating robotics, as we find ourselves on the tipping point between covering such liability under established civil rules, and the need to develop new rules to cover acts for which the direct involvement of a human actor will be to a great extent eliminated e.g. a robot that takes autonomous decisions through machine learning beyond its original programming.
These are not new issues being raised by the Parliament. Indeed, the European Commission has already launched a consultation into the adequacy of Directive 85/374/EC on Liability for Defective Products, specifically citing questions as to whether the existing rules are fit for purpose in relation to “technological developments such as the Internet of Things and autonomous systems”. Questions of the adequacy of existing rules of liability for new technologies have also been considered in the work of Working Group 4 within the Association for the Internet of Things Innovation (AIOTA), and dealt with in the recently-published AIOTA report.
Whilst questions of liability for damage are important, and potential challenging when it comes to smart robotic machines, the more fundamental question in the European space is how to regulate the safety and performance of such machines as a precondition to them entering the market in the first place. It is at that level that questions of who is responsible, and for what, will bite the hardest in the short term, and which will potentially represent the greatest threat to innovation if not handled correctly.
Safety rules
Besides, all products – including robots – generally have to be safe, i.e. complying with reasonable safety expectations by meeting state-of-the-art requirements. However, in an area that drives developments at a rapid speed and also considering that robots’ and AIs’ abilities will probably be expected to surpass human abilities, it will be increasingly difficult to define safety standards. Hardware developers and software programmers are increasingly seeking for legal guidance to determine the requirements they are working against.
Social and ethical considerations – and a new agency
The European Parliament’s rapporteur in her draft report had suggested taking account of possible negative consequences on the job market that the use of robots might result in. Interestingly enough, Bill Gates recently argued that robots that take over jobs should in fact pay a form of income tax.
However, the Parliament’s majority did not agree to include the rapporteur’s concerns regarding the impact on the workforce in the final text approved. Instead, the Parliament potentially created a new workforce by asking the Commission to consider setting up a European agency for robotics and artificial intelligence, to supply public authorities with technical, ethical and regulatory expertise.
In the field of robotics and AI, ethical considerations have always been a driving factor. As an example, already in 1942, science fiction author Isaac Asimov defined “the tree laws of robotics”, e.g. that a robot may not injure a human being. In doing so, it does not come as a surprise that the Parliament appreciated that the growing use of robotics raises ethical issues, for example, to do with privacy and safety, and proposed that a voluntary ethical code of conduct on robotics for researchers and designers is created, to ensure that they operate in accordance with legal and ethical standards and that robot design and use respect human dignity. The proposed charter could have varying binding effects as it would constitute an umbrella framework, a soft law instrument, covering practices such as the robotics engineers’ ethical code of conduct, the conduct for research ethics boards, and the designer licence and user licence.
Conclusion
With many countries such as US, Japan, China and South Korea, already developing their own rules on robotics and artificial intelligence, the EU is now aiming at taking pre-emptive steps to ensure that any vast roll-out of robotics does not fall below its highly established standards when implemented in the EU.
Such rules will inflame the necessary discussion of what happens next, once such robots become part of our daily lives. Developers, manufacturers and users of robots should closely follow this important debate which is likely to require a steep learning curve for the legislators and the enforcing authorities. It is an area that calls for a flexible and forward-looking approach to law making and regulation, to avoid a legal environment which becomes characterised by inefficiencies, stifled innovation, wasted opportunities, and the need for constant amendments as these technologies present new challenges that seemed to previously only factor into sci-fi novels.
