Opinion & Analysis

Agreement or Polarization? How European Institutions and MEPs React to Regulating AI

Following pressures to create a legal framework for artificial intelligence, European institutions are starting to draft initiatives and proposals.

So, let’s take the next step in our AI journey and look at how European political actors are positioning themselves today. Spoiler alert: with caution and slight optimism.

We start by looking at the latest announcement from the European Commission, as it sets the tone and direction for the entire European Union.

We then go to our main course and specialty: the voting behaviors and opinions of Members of the European Parliament on issues surrounding AI. These are likely to reflect the positions of governments and the European society and to predict trends in decision-making and regulations. As with the other policy areas, the political forces at the EU level have very different views on  the regulation of AI, as some decision-makers enthusiastically embrace the new technological developments, while others are more concerned.

We finish with some food for thought that politicians might have trouble digesting. But it hopefully encourages everyone to consider the unlikely, as it’s likely to occur with the unpredictable rise of AI.

Investments in AI, a European Priority

AI is no longer science fiction and EU institutions and Member States must take measures today to keep up with the new technologies. This was the main message of the European Commission two weeks ago, when it released a set of guidelines and measures meant to boost investments and help create a legal and ethical framework for artificial intelligence by 2020. The measures announced include:

increasing the investment to €1.5 billion for the period 2018-2020 under the Horizon 2020 research and innovation programme;

mobilizing the European Fund for Strategic Investments to provide companies and start-ups with additional support to invest in AI;

proposing legislation to open more data for re-use and measures to make data sharing easier;

creating an “AI-on-demand platform” that will provide access to relevant AI resources in the EU for all users.

With these measures, the Commission hopes to “support the development of AI in key sectors, from transport to health; to connect and strengthen AI research centers across Europe and encourage testing and experimentation.” It also hopes to catch up to the US and Asia that are ahead of Europe in this respect, analysts say.

EU Member States Team Up for AI

Just weeks before the Commission’s announcement, 24 EU member states signed a joint declaration vowing to work together to deal with the good and bad of the rise of artificial intelligence. The signatories promised to favour investments in AI research and to modernize national policies, all under a European approach.

Members states are teaming up. The European Commission is proposing clear measures and guidelines. In 2017, the European Parliament adopted a resolution for the creation of a European Agency for Robotics and AI. With everyone on board, what could go wrong?

First, the rise of artificial intelligence is complex. There are many aspects to consider and different speeds at which the European Union can move. Member States are likely to agree on general matters but disagree on the details – it wouldn’t be the first time. Second, we have already noticed conflicting opinions and voting behaviors in the ranks of the European Parliament – which likely mirror the same trends at the level of member states governments and the society overall. The paragraphs below provide a overview of the positions of the political forces in the EP on the hottest issues concerning the rise of AI.

EU Registration System for AI

To better track advanced robots in the European Union, 66% of EU Parliamentarians supported the idea of introducing a Union system for the registration of advanced robots, which would be managed by a designated EU agency for robotics and artificial intelligence (EU Agency for Robotics and Artificial Intelligence). This proposal is backed by the two biggest parliamentary groups in the EP, the centre-right EPP and centre-left S&D, with help from the Greens and the Italian 5 Star Movement.

However, the other political groups are not concealing their scepticism over such a proposal. The far-left GUE-NGL, centrist ALDE, right-wing ECR and far-right ENF are all against this proposed centralization of controls on advanced robots. While some of Eurosceptic groups are concerned on the extension of EU powers on such a sensitive policy area, other members are worried about over-regulation of AI, as this could potentially stifle innovation.

Although in May 2017 the European Commission rejected the Parliament’s suggestion for creating such an agency, the political dynamics have been evolving over the last year.

Robots Replacing Humans

Automation is already threatening the job of millions across the world. While predictions range from optimistic to pessimistic, institutions must be ready to deal with the changing dynamics of the workforce and workplace.

Robots might replace humans in a wide range of jobs, from operating vehicles and machinery, to doing routine surgeries in hospitals. While we’re ready to accept concepts as self-driving trucks, what about jobs that require a so-called “human touch”?

If, for example, one day robots could carry out our primary caring jobs in hospitals and caring centers, would we allow them to do so? Almost four fifths of MEPs (79%) would not allow this to happen, thinking that robots will dehumanize caring jobs no matter how much they look and behave like human beings. On the other hand, most MEPs from ALDE and ECR think robots should get a shot at performing caring tasks.

The consensus between EU law makers (87%) seems to be that robots should complement humans, but not replace them. Among the political groups only ALDE embraces this rather equal relationship between humans and robots in the future. ALDE’s views are also shared by Swedish right-wing MEPs from both the Moderate Party (EPP) and Sweden Democrats (EFDD).

If robots cannot replace humans entirely, how about just a part of our body?76% of MEPs welcome the potentialities for human enhancement through cyber-physical systems. However, a conservative front made up of ECR (Poland’s PiS and UK’s Conservatives), ENF (France’s National Front and Italy’s League), as well as most Croatian, Czech and Italian EPP members, is against exploring this possibility. Interestingly, most French Greens agree with the conservative forces, as French MEP José Bové warns against initiatives that could pave the way to transhumanism.

To regulate the use of human repair and enhancement technology and decide on the intricate ethical problems that might come along, most MEPs support the creation of committees on robot ethics in health care institutions (72%). However, most ECR members are against the establishment of such committees, so do ALDE and the Swedish right-wing members (from the Moderate Party and the Swedish Democrats). While the most conservative members oppose the initiative because they are afraid these committees could stimulate discussions on human enhancement, the opposing progressive parties are afraid ethical committees could slow down the developments of cyber-physical systems.

New Welfare Models

It’s not a question of if, but a question of when and at which speed robots and AI will take over many of the activities humans perform today. To prepare for this likely future, a comprehensive assessment on the impact of robots should be undertaken, especially on employment. 89% of MEPs support such a call, while only the members of centrist ALDE, Swedish conservatives (from EFDD and EPP), and the Dutch far right do not find it necessary.

What about labor tax? Should a tax on robots be introduced to make up for the financial loss that governments are likely to face because of automation? For the time being, MEPs seem divided along a left-right axis (with some exceptions). 47% are in favour of a tax on robots (S&D, Greens/EFA, GUE-NGL, Italy’s 5 Star Movement and the Freedom Party of Austria), while 49% are against it (ALDE, EPP and ECR).

MEPs were similarly divided on the possibility of introducing a general basic income if the development of robotics and AI were to lead to drastic changes in the current social security system of the Member States. Leftist political groups (S&D, Greens/EFA, GUE-NGL, Italy’s 5 Star Movement from the EFDD group and Dutch D66 from ALDE) would back such a proposal while ALDE, EPP and ECR are against it. Overall 53% of MEPs reject the introduction of a general basic income.

Finally, 50% (against 49%) of MEPs are afraid that the development of robotics may lead to a concentration of wealth and influence in the hands of a few, just as internet technology has created digital giants like Google, Facebook and Amazon. In this case, far-right ENF and some Italian, Luxembourgish and Croatian EPP members have joined the left-wing forces in voicing their concern. However, the optimistic front can count on most ALDE, EPP and ECR members, as centrist and right-leaning MEPs are less worried about this trend.

Legal Status and Legal Liability for Robots

In our previous article, we discussed some of the questions that are emerging around granting rights to artificial intelligence entities. More than half of MEPs (51%) agree on the need to introduce a legal status of electronic persons for robots in the long-run, without giving details about how this legal status would look like. 46% of MEPs oppose it (Greens/EFA, the Italian Democratic Party from S&D, ALDE, ECR and ENF), while the EPP group is split on the issue. The majority of EPP members are in favour of creating a legal status for robots, but the Croatian, Czech, Dutch and Spanish EPP delegations, as well as about half of the members of German CDU, voted against such an initiative.

As far as legal liability for the actions of autonomous robots is concerned, most MEPs (74%) agree on the need of new frameworks to make sense of the legal liability of various actors regarding the acts of autonomous robots. Still, ALDE and ECR groups, and some Italian, German and Croatian members of the EPP, as well as the Italian League and Dutch Party for Freedom (ENF), oppose the introduction of a new framework to define the liability for the acts and omissions of robots.

Direct Democracy, Powered by Machine Learning

While politicians are taking the first steps towards creating a legal framework to regulate AI, there’s already talk about bypassing the political process altogether.

According to FastCompany.com, MIT Media Lab’s Cesar Hidalgo argued in a recent TED Talk that politicians are failing at representing their constituents and that our current representative democracy could be replaced by a system in which everyone votes on every issue. And don’t worry about having to take too many decisions. Hidalgo has the solution: an algorithm can learn your political preferences based on your daily behavior and take decision for you.

Does it sound too out there? Sure. Now it does. But how about in five or ten years? Just remember that self-driving cars were a sci-fi dream a short while back. Politicians, as all other citizens, might have to rethink their roles too.