MEPs have approved rules for keeping humans firmly in charge of Artificial Intelligence (AI). They include ways to establish liability in law, for example where driverless cars cause accidents. They have also called for ethical standards to be built in to AI algorithms and robots that work for humans, and standardisation across Member States to ensure a level playing field for technology companies.
After a heated debate, the vote was passed on Thursday by a large majority, 369 to 123 with 85 abstentions. But clauses that would introduce a basic state allowance for people who lose their jobs to robotisation were defeated. A coalition of right-wing parties voted them out. The Rapporteur who steered the new policy, Luxembourg socialist MEP Mady Delvaux Stehres, expressed disappointment. She fears digitisation will create mass unemployment: in areas like the service industries, where call centres are already automated and in transport and logistics through driverless vehicles and drones. According to the European Commission’s own statisticians, Slovakia and Germany head the list of countries where most jobs are at risk from automation (55%), with Russia at the bottom (30%). The jobs most at risk are those that require no social skills nor advanced mathematical competence, as shown here:
Social Skills and Change in Share of Jobs in the United States, 1980 to 2012
Now the European Commission can choose to bring in directives that will enforce these new Civil Law on Robotics regulations across the EU. The people of Europe can also have a say: Parliament’s Legal Affairs committee has opened an online public consultation that lasts until the end of April 2017. This new law is the result of two years of research and consultation. Netopia started this debate, back in 2014, with the publication of its research report “Can we make the digital world ethical?”.
Netopia launched the findings in Brussels with senior EU Digital Agenda advisor Nicole Dewandre, and also presented them to the French Senate in Paris. The report recommends “human primacy”. That means that robots and Artificial Intelligence should always be programmed to put people first. It calls for ethical codes to be built in to software, and for “device sanctity” which means that a computer, tablet or smartphone – and the data on it – is regarded as part of its owner’s legal person, so that it cannot be violated.
This is an historic moment because it means that politicians have realised the profound impact that technology now has on our lives and they are moving to regulate that finally politicians have begun to grasp the ramifications of the new digital world.
In my interview with her, Rapporteur Mady Delvaux Sèhres insists that MEPs are not trying to stop technological advances nor stifle innovation: “The European Parliament thinks there should be ethics by design… I understand that it will be difficult and time consuming but Parliament hopes by standardisation to prevent unethical robots from coming to the market”
Delvaux’s committee has developed ethical principles relating to human rights to safety, privacy, integrity, dignity, autonomy and data ownership
They include creating legal liability and insurance for driverless vehicles and compensation for victims when they go wrong. And there is a perceived need for ‘health warnings’ on humanoid robots, such as those already widely used in Japan to care for elderly people, replace sex workers and even to staff the world’s first robot hotel. “We must understand that they are just very, very smart machines. They are programmed to smile at you and you must not fall in love with them” warns Delvaux.
British MEP Mary Honeyball, a member of the Legal Affairs Committee, is enthusiastic in her support of this initiative. But she regrets that the UK will miss out on the benefits of the new regulatory system because of the referendum vote to leave the EU: “We must not lose sight of the fact the UK will not be part of this exciting European discussion and debate as it develops, and this is another example of the folly of us voting in favour of Brexit.”
The next step in the process is the public consultation. It aims to raise awareness of AIs in everyday life, for those of us who think that humanoid robots are cute and our AIs – Siri, Alexa and other digital daily helpers – are our friends. Meanwhile technologists and lawyers need to tackle the pressing need for law and ethics to catch up with the pace of technological change. MEPs have issued a challenge, and the industry must respond.
More on Artificial Intelligence:
Peter Warren, a researcher into digital ethics since the late 1980s as the editor of Future Intelligence is arranging a conference on the issues raised by the technology with the international law firm Cooley and the Institution of Engineering Technology, on 25th April 2017. This is one of a number of conferences that will debate the topic, with the oldest running AI conference Artificial Intelligence and the forth coming Simulation of Behaviour making its theme ‘Society with AI’ at Bath.