Mady Delvaux-Stehres (European Parliament): "There is no way a robot should be treated like a human being"

LEADERS LEAGUE. What types of machines does this report cover?

Mady Delvaux-Stehres. This is a central question since a definition will determine also the scope of the legislation. This topic touches so many different areas such as the medical sector, industry, defense, finance, education, daily home robots etc. If a vacuum cleaner, an autonomous car or an algorithm can all be considered robots, how to cover them in one definition? Should they be treated under the same legislation? This report seeks to cover the range of “smart autonomous robots” by taking into account today´s and tomorrow's technological progress in the area of robotics and artificial intelligence, it will be the duty of the Commission to clearly define what is a robot. This report identifies different criterion which could shape the definition: Degree of autonomy, physical support, self-learning, adaptable behavior to the environment etc. Since there hasn't been a proposal so far for a legal definition of this product, we want to emphasize the essential need of such.

What today are the main legal issues posed by robots? How urgent is it to legislate them?

I identify five areas in which we should intervene: liability, safety and standardization, data protection, ethics and education. We are already behind. When my colleagues and I started the working group four years ago, we thought we would have time, but the technology is moving so fast, that we cannot wait anymore. I really would have wished the Commission would make a legislative proposal for this legislature ending in one year. But for except an announcement planned for April 25th, nothing further will be done.

What would be the consequence if each member-state set up its own rules?

We need a single piece of EU legislation to develop a coherent and safe market for robots. We need to give legal certainty and support to our companies and protection and safety to our citizens. If the market is fragmented by different national legislation, it will not be possible. Can you imagine a car company producing different cars in response to different legal criteria? Can you imagine a user stopping his autonomous car at the border? This revolution is a great opportunity for the EU market but if we don’t act now, we will miss the turning point.

We have to decide what we want to do with this opportunity or the rest of the world (US, China, etc.) will decide for us.

Germany already has legislation on autonomous cars, France started the legislative work with the Villani report on AI and robotics, the UK House of Commons also gave its report with guidelines for legislation, a law on liability is working its way through the legislature in Czech Republic etc. The EU has to take the lead to be stronger in our globalized world.

Should we accord robots the legal status of humans, corporations or something else?

There is no way a robot should be treated like a human. A robot is and will remain a machine. In our report, we proposed to study, among other possibilities, the use of an e-personality which would function a little bit like the legal personality of companies. Indeed, several experts, including professor Bertolini of the Sant’Anna School of Advanced Studies in Pisa, and prominent Paris-based new technologies lawyer Alain Bensoussan have proposed that, where an accident with a robot occurs and due to the degree of autonomy of the robot no one can be held liable, the victim could be compensated by using this e-personality to raise a fund for victims. This possibility has come in for criticism, firstly, because there has been misunderstanding of the legal concept leading to the assessment that a robot would be treated like a human (which is of course not the goal at all).

Secondly, consumers’ associations and experts have pointed out that this could mean companies are no longer held accountable for the ‘actions’ of their products and might be tempted to take bigger risks.

How do you believe the liability of these robots should be framed going forward?

This has to be studied carefully. So far, my favorite option is the risk-management approach. Meaning the person who was best placed to prevent the risk should be liable. I also recommend in my report the creation of a mandatory insurance scheme to ensure compensation for victims without stifling innovation.

What measures, specifically, are being taken to regulate robotics and AI?

In the US, it is a legal race between states to be the most welcoming for tests. Unfortunately, this sometimes leads to too much risk, and tragic consequences, such as the Uber accident killing a person in March 2018, have resulted. In China, there is massive investment in
the most strategic technologies and in providing the best place for production. Japan and South Korea have passed legislation on the standardization and safety of products.

Compared to other countries, would you say the European Union is playing catch-up on this issue?

The EU has produced great engineers, creators and developers, but so far we have not been able to hang on to them. No country in the EU does has a giant investing massively in research, so the EU should take on this duty.

A coherent and safe legal framework would be an asset but we have to invest much more to help our creators.

I believe that our ethical approach will also bring more confidence in our products. In the report, we also underline the educational challenge and the transition in the employment market as it is not only about economy, it is also about our society and social cohesion. This opportunity should not be missed. It’s time to decide which future we want.

Interview by Camille Guével

Find more analysis, articles & interviews in our 2018 Innovation, Technology & IP Report to be released in May.