We are in the middle of a cognitive revolution. Germany must be more ambitious with its strategy for artificial intelligence and promote its implementation as a national priority. Otherwise the rules of this new age will be written – yet again – in Silicon Valley or Shenzhen.
Increasing computing power, breakthroughs in research and growing data pools: Artificial Intelligence (AI) is developing rapidly and increasingly automates cognitive processes. Decreasing costs for technology components, the almost ubiquitous availability of computing power through the cloud, open online courses for learning to develop AI and a rising number of open data initiatives allow everyone with internet access to participate in it and shape it. At least in theory. In practice, AI is being increasingly monopolised and economic outlooks show that the benefits of AI are unequally distributed across the globe: By 2030, 26 and 14.5 percent of AI powered GDP growth in China and the USA respectively. As the rapid progress in AI research and applications redefines the balance of economic and political power worldwide, countries across the world have been embracing AI with ambitious national strategies, in an attempt to catch up with the AI superpowers and positioning them at the forefront in this field.
The Obama administration was the first to lift AI to the status of a national priority in 2016 based on which the current administration has started to develop a comprehensive strategy. This triggered what some observers call an AI “arms race”at the global level. So far, the US holds the pole position: No other country in the world graduates more PhD candidates in AI and related fields every year than the USA, an estimated 3,000. In Germany the figure is not more than 170 yearly. The USA is also home to most AI startups globally, roughly 1400, compare to Germany’s 100. Similarly, almost three-quarters of all internationally enforceable AI patents are filed by US-based actors while German ones only contribute two percent so far. With recent advances in Quantum Computing, like the IBM’s launch of world’s first commercial quantum computer, the US companies appear to be cementing their leading position in cognitive computing. In turn, the fertile US ecosystem of excellent research institutions, innovative entrepreneurs, experienced investors and agile public actors is likely to keep attracting even more top talent, capital and intellectual property.
Only China can compete with the AI superpower USA. In 2018, nine of the world’s twenty most valuable technology companies came from the Kingdom of the Middle, eleven from the USA. Apart from the huge domestic data pool of its over 800 million internet users, the country accounts for more than 200 of the Top-500 supercomputers. The pressure on Chinese entrepreneurs to survive in an extremely competitive environment in which the copying of business models and products is culturally accepted, has led to the growth of companies whose innovative strength partly eclipses that of American tech groups. WhatsApp is now learning from the functions of its Chinese counterpart WeChat and not vice versa. The Chinese government sees artificial intelligence as a strategic infrastructure project for its leadership in the 21st century. The city of Tianjin alone has announced investments of 12 billion euros for AI. The relative weakness of domestic science is offset by Chinese corporate technology groups that invest billions of dollars in establishing research and development centers in the USA, Europe and neighboring Asian countries.
Other Asian countries have also recognised AI as a strategic national asset. Japan’s government has declared the fusion of the physical and cyber worlds to be a national goal in its concept of the “Society 5.0.” Here, too, AI is regarded as an indispensable prerequisite for the transformation to a cognitive society. The AI-powered “Society 5.0” is not only an economic concept to prop up growth in light of an ageing and shrinking society, but has social dimensions as well through, for instance, the use of robots in elder care. Neighboring South Korea will invest 1.7 billion euros in AI by 2022 and one billion euros in brain research. And Singapore, encourages industry, researchers and citizens to experiment and build “AI for Everyone” and is at the forefront of AI Governance in South East Asia.
Meanwhile, Canada, India and the United Arab Emirates aim to establish themselves as important hubs within global networks for science, data pools and applications. In Russia and Israel, as in the US and China, the military is an important driver of AI development. These defense roots of AI is less about ‘killer robots’ that pop up so often in Western media, but more about AI-enabled mass surveillance, cyber warfare and cybersecurity.
And Germany, the worlds 4th largest economy? The country has two strategic assets: a good starting position in research and a globally networked economy. In order to build on this and participate in the cognitive revolution, the German government published its own AI strategy in December 2018. It intends to allocate a budget of three billion euros to 2025. However, in March 2019, German newspaper Handelsblatt reported that this amount would only entail EUR 500 million of new funds while the rest is taken from existing budgets. This could be the mistake of the century! The same newspaper reports that the ministries responsible for the implementation of the strategy have not yet put forward comprehensive measures for this fiscal year. But even if the full three billion euros were available, it is questionable whether this would suffice: Although this number exceeds the funding provided by other European countries, such as France or Great Britain, it is small relative to Germany’s global economic weight and the aspiration to establish the brand “AI Made in Germany” and a European “third way” vis-à-vis the US and Chinese models. The fact that the German national AI strategy comes late also does not help, because time is of the essence. After all, the development of a digitally competent and responsible civil society requires time for skill development and experimentation with new models of working, learning and earning across many different sectors in society.
In order to be able to implement the German strategy effectively and to help shape it on the European and global stage, Germany has to focus on and boost three aspects of its strategy in particular:
1. An Apollo-like project for AI
Germany is the world champion in talking technologies down even before they have ever been used. A constructive and courageous openness toward new technologies would be desirable. Ensuring that artificial intelligence does not remain a purely economic topic – as the term “Industry 4.0” implies – requires a broad vision rooted in thorough definition of what AI actually is and what it is not. This should reflect people’s concerns, but it must also ask questions about the future that explore the technologies’ potential: What could opportunities for enrichment, health and prosperity in everyday life through artificial intelligence look like? How can we improve society and make it more equitable with the help of AI? How do we bring people with us on this journey? This requires more than just a communication campaign, as currently envisaged in the AI strategy. Young people in particular need to be enthusiastic about artificial intelligence, be allowed to experiment with it and learn. This is the only way they will be able to set up companies in this field at a later date. We need to ignite a Sputnik or Apollo spirit in and with this generation.
This openness also means not excluding entire fields of research, such Artificial General Intelligence, because we feel too threatened too soon. Such an intelligent machine would possibly be superior to humans in several fields and, yes, that may sound scary at first. But not developing a smart partner for humans could be a missed opportunity. The question arises as to why a country that is participating in a supercollider that makes it possible to study the universe on the smallest particle basis would not want to develop a “super thinker” that integrates artificial intelligence, neurology, brain research and brain-computer interfaces, which could substantially improve our understanding of the human brain. A prerequisite for such a “super thinker” is the development of the quantum computer, which should not be left to the AI super powers. Germany already has a few players that are exploring this field, such as Volkswagen and Carl Zeiss AG, whose research and development should be integrated and specifically promoted. With a CERN-like European Center for Cognitive Computing, Germany could strengthen its complementary position in the global research landscape for artificial intelligence in cooperation with other European countries. Hardly any other country besides the US has set targets in this area. Let us recognise the transformative opportunities, not just the risks!
2. A central steering structure to ensure speedy implementation
To achieve goals that are more visionary, a central steering committee should be established. It would monitor the implementation of the strategic objectives between ministries and the AI ecosystem and adjust them if necessary. This will be particularly necessary against the background of rapid technological advances. Various ministries are already working on this, supported by the Digital Council of the Chancellor, the Minister of State for Digitization and the newly founded “Agency for Leap Innovation”.
While it is encouraging that the federal government has recognised artificial intelligence as a cross-sectoral issue, the implementation of the strategy threatens to become a bureaucratic and therefore slow process. When it comes to artificial intelligence, Germany cannot afford to go at the same slow-motion speed as it showed in the case of electric and autonomous driving or strategic infrastructure projects such as magnetic levitation railways or the new Berlin airport. A steering committee would have to be equipped with the freedom to make quick decisions, even without the involvement of all responsible state actors. This may also require giving the federal level more decision rights to more effectively steer the implementation of the national AI strategy across the involved actors.
3. Leadership in ethical AI by orchestrating a global Digital Magna Carta
Values and ethics questions around the use of AI are important. However, Germany should not discuss these merely on a national level only and in a variety of expert committees, as is currently the case with the Federal Government’s Data Ethics Commission and the Bundestag’s Enquete Commission. The German AI strategy does call for a global debate in so-called AI observatories and in cooperation with the United Nations, the G7 and G20 states. However, these organizations neither allow civil society to fully participate nor are they particularly adept in technical matters.
AI ethic principles and governance initiatives from industry bodies, civil society and state actors are blossoming across the globe, like the EU Commission’s Ethics Guidelines for Trustworthy AI, OECD’s endorsement of AI principles, Singapore’s Model AI Governance Framework, France and Canada’s International Panel on AI. Yet, they usually do not span across regions and have no concrete mechanisms for implementation. But this is crucial, as research shows that AI ethic principles do not translate in ethical decision making in software development. Germany could set itself apart from other countries in building an alliance for developing ethical, human-centric AI governance and safeguard mechanisms for bold transformation projects.
The German government could join this initiative and advocate for a global Digital Magna Carta as described in our book “Solomon’s Code” and also called for by the inventor of the Internet Tim Berners-Lee, as well as Anthony Giddens (British House of Lords). Germany should use its full diplomatic weight and the strategic assets of its multilateral and value-based foreign policy to this end and also involve countries of the global South, as provided for in the AI Strategy. Germany’s historically pragmatic focus on rules and mechanisms for a “social market economy” can serve as a credible precedent for hands-on rule-making that fills the notion of a human-centric “third way” with substance and ensure, that ethic principles do not remain paper exercises. Otherwise, the rules for the Cognitive Revolution will be written in Silicon Valley or Shenzhen, just as they were for the Information Revolution.
So, dare more AI, Germany!
About the Authors
Dr. Olaf Groth serves as Professor for Global Strategy, Innovation and Economics at HULT International Business School and program Director for Digital Futures at HULT-Ashridge Executive Education. He is Managing Director of advisory thinktank Cambrian Group, Visiting Scholar at UC Berkeley’s Roundtable for the International Economy, a member of the Global Expert Network at the World Economic Forum and co-author (with Dr. Mark Nitzberg) of the new book Solomon’s Code: Humanity in a World of Thinking Machines. Previously, Dr. Groth was an executive with hi-tech multinationals.
Tobias Straube is Principal at Cambrian Group and led with Olaf Groth the development of a study comparing national AI strategies for the Konrad Adenauer Foundation, and worked for eight years for the Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ), the Friedrich Naumann Foundation, and Bündnis 90/Die Grünen in Germany, Latin America, Africa, and Southeast Asia to develop and implement regional governance, higher education, and entrepreneurship initiatives. Straube lives in Frankfurt. This article reflects his personal opinion.
Toni Kaatz-Dubberke supported the comparative study on national AI strategies for the Konrad Adenauer Foundation as a Senior Analyst at Cambrian Group. For eight years, he has been active in international cooperation projects in Asia and Africa, at present for the Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ) GmbH in Germany. Since 2005, he also works as a freelance journalist. This article reflects his personal opinion.