Robo go Germany gets smart about Artificial Intelligence

Berlin is playing catchup on AI. That may be a good thing if it avoids mistakes, like an excessively narrow focus on national interests. Germany should back a European AI policy, writes a Mercator Fellow and Economist author.
Quelle: Moment/Getty Images
Berlin finally joins the robo highway.
(Source: Moment/Getty Images)

Germany has its own national artificial-intelligence (AI) strategy at last. And at first glance, the long wait was worth it. As a “smart follower,” the government in Berlin was able to learn from other, earlier plans. The German paper is more than a laundry list of new research projects. It covers a lot of important conceptual ground, for instance the best way to organise access to data, the key ingredient for most AI services.

Yet nothing is perfect. Critics rightly point out that ethical questions are given short shrift, such as the problem of preventing algorithms from discriminating against certain groups. But an even more important lacuna may be the strategy’s extremely national outlook. What role AI will play in foreign policy, and how it will change international relations, is only mentioned at the margins.

Sadly, Germany is in good company here. Most countries increasingly see AI through their own national lens. Experts are already speaking of an emerging “AI nationalism,” with countries mostly following their own techno-political interests. Among other things, governments block data flows and protect national AI champions. The bigger countries are also taking the first steps toward starting an AI arms race.

The reason for this is the huge economic impact that most expect AI to have. The technology is best understood as a form of “collective intelligence.” Acting as digital distilleries, big data centres extract patterns from data, most generated by individuals. These patterns are then turned into all kinds of services, from facial recognition to self-driving cars, that will automate much of the physical world.

Because AI is likely to be injected injected into everything and anything, it is often compared to electricity. Whether this is the best metaphor remains to be seen, but the technology is likely to boost growth in a similar way to power a century ago. According to one estimate, by 2030 AI will increase global GDP by $16 trillion (€14.1 trillion), about five times the current size of Germany’s economy.

There will be losers, as well as winners, however. AI is likely to shake the world out of its current balance. Because they generate plenty of data, China, and to some extent even Russia, have a chance to challenge America economically and militarily. AI could even alter the relative competitiveness of democracy and dictatorship. It lets governments centrally process huge amounts of data – something that previously escaped authoritarian regimes and eventually led to their failure.

With so much at stake, it should not come as a surprise that countries focus on their own interests. Some experts argue that AI nationalism will fade once the technology has matured and winners and losers emerge. But the current conflict over trade between America and China points to an entirely different scenario: that the world will fall into two AI camps, with smaller countries having to decide which side they are on, a bit like during the Cold War.

How should Germany react to such a development? For one, it must build a strong AI industry of its own. But it can only achieve this goal if it co-operates with its European partners. The country’s data pools, for instance, are much too small for it to be able to compete with the AI superpowers on its own. This means Germany needs to more closely co-ordinate its AI policies with the European Union.

Secondly, together with the EU, Germany must push for the main building blocks of AI to remain global public goods. Currently, most AI research quickly finds its way into the public domain. To avoid the drying up of this AI commons, Berlin and Brussels should actively support organisations that work on creating AI standards and preventing the technology from being controlled by governments and companies.

Thirdly, if Europe really wants to stand for a “third way” in AI – between a Chinese surveillance capitalism and the American monopoly variety – a goal many in Berlin and Brussels support, it has to quickly fill this rather vacuous concept with meaning. The EU’s new General Data Protection Regulation, whose principles are being copied around the world, can only be the start.

Similarly, Germany’s national AI strategy can only be a first version, which must be quickly followed by versions 2.0 and 3.0. “AI policy will become the single most important area of government policy,” writes Ian Hogarth, an AI expert and entrepreneur, in a widely read online essay. This, of course, is an exaggeration – but not by much.

To contact the author: [email protected]