Refining the rules on robotics16/02/17
The European Parliament is to vote today (16 February) on the legislative initiative report regarding the civil law rules for robotics. MEP and rapporteur of the text Mady Delvaux explained all to PEN
Robotics and Artificial Intelligence (AI) is a term that implies the use of a computer to model and/or replicate intelligent behaviour. From medical diagnosis and equipment to military planning and drones, robots are not only growing in demand but are also becoming a part of our everyday lives.
As the use of robotics expands, so does the international debate surrounding the need to advance the understanding of the legal, ethical, societal and operational implications that this level of intelligence demands.
What legislative challenges do you anticipate when it comes to international co-operation/competition – including the US, China, South Korea and Europe, etc.?
Firstly, I would like to point out that the European Parliament is the first institution in the world to work on a legislative framework for robotics.
Robotics is a significant sector in the EU and Europe is one of the world-leading regions in industrial robotics with a share of more than 25% of supply and use.
According to the European Commission, the EU accounts “for around a quarter of the global production in industrial robotics and a 50% market share in professional service robotics”. The EU is a leader in this sector and the harmonisation of rules at this level would bring much more legal certainty.
On the one hand, certification would guarantee to EU citizens that robots comply with safety, health and environmental rules. On the other hand, it would also bring predictability and legal clarity for the industry. A legal framework would indeed foster a certain level of competitiveness, as companies would be aware of what they must or must not do, which will make predictability better in terms of success.
The report highlights the fallibility of Asimov’s laws. What is needed to ensure (human) safety and where should this be sought?
If we reached harmonisation with the technical rules, the companies would need to comply with a certain number of rules, which will in turn protect the consumers. We also recommend the proposal of a charter for engineers in the annex, so that they are permitted to programme the machines in the human beings’ interest.
If we talk about safety in the sense of respecting private life, the report calls for transparency and compliance with the EU directives.
Transparency is linked to civil liability. If a person has an accident because of a robot, the victim should be compensated.
However, it is important to identify who is responsible. We would therefore ask that robots are provided with ‘black boxes’ at industry level to ensure that the algorithms are readable and to understand why a robot would make such a decision and who can be held liable for the damage.
It is fundamental to respect the human beings. They should be the winners of robotisation in society.
When it comes to the classification of robots, are existing legal categories sufficient? What challenges are there in developing new ones?
The classification of robots is the ideal. This is the first question we had and it is hard to answer.
Today there are robots for every purpose and in every field, such as the autonomous cars or medical robots.
We wondered if non-physical robots, such as the algorithms that calculate investment in the stock exchange, can be defined as ‘robots’ and as such was have decided not to label them as robots, as this would be too complicated and broad.
The main point is to follow the development of robots. We should then have a definition that is neither too broad nor too narrow. That is why we have not identified a clear definition of robots and we have called upon the commission to conduct studies into acquiring one. Nonetheless, it is obvious that medical robots and autonomous cars cannot be governed by the same rules.
We aim to determine a definition which will protect the consumers as well as foster innovation. We have therefore left the door open.
The report calls for the establishment of a ‘European Agency for Robotics and AI’. Why is this important?
Robotics is constantly moving and going further. We need a stable and also flexible legal framework. We think a European agency is essential to deal with these specific questions. Robotics is a topic that requires expertise and high-level competence. As robots will be used across Europe, the rules and guidelines should be harmonised at the EU level. This is why I am convinced a European agency should be set up to face the coming challenges.
The benefit of a European agency for EU citizens and industries would probably largely cover its cost.