Numerous countries like China, France and Japan have declared Artificial Intelligence (AI) a key technology and announced comprehensive plans to promote research and development in AI. The German government has also begun to work on an AI strategy and just published a first blueprint with the core themes. In this paper, we argue that such a strategy needs to be broad and comprehensive, focusing on the development of an internationally-competitive AI ecosystem in Germany. A strong AI ecosystem is characterized by strong networks between science, economic actors (big companies and startups alike) and society at large. Innovations arise in particular from close exchanges and collaboration between researchers, developers, universities, companies, investors and startups. To promote such an ecosystem, a wide range of different political measures on different levels have to be integrated into a broader, comprehensive strategy. This paper discusses the central building blocks of an AI ecosystem in Germany, and offers concrete ideas and recommendations for an AI strategy based on this ecosystem approach. 1. AI research: Compared to other countries, Germany is lagging behind in research expenditures and must drastically increase them. Research support needs to be open to different technological approaches within AI. It also needs to be more agile to better react to emerging trends and new opportunities in AI research. Better work conditions overall are needed to compete for the best AI talent worldwide as well as clear benchmarks to measure progress in AI research. 2. Development of AI competencies across society: We do not only need top research. We also need broadly distributed AI competencies in society. Thus AI should not only be taught in computer sciences, but core AI modules should also be integrated into engineering and natural science programs, and be taught at schools of applied sciences. 3. Data as a basic resource for AI development: A strong AI ecosystem needs data for research and for the development of AI applications in industry, particularly with regard to deep learning (DL). This dimension of the ecosystem needs far more attention in Germany. Possible approaches to mobilize data for AI include the development of data pools and more advanced methods of anonymizing or synthesizing data. It is hard to compete with the big Internet platforms from the United States and China in terms of quantity of data. Instead, special emphasis on machine data, quality of data and alternative approaches to AI that can work with little data could be the cornerstones of an alternative path to a strong AI ecosystem. 4. Infrastructure demands for AI: Deep learning requires not only huge amounts of data but also great computing power. A national AI strategy would address the question of how we can ensure middle- and long-term access to the most powerful processing hardware possible for German AI research and applications. 5. AI development and AI application in the economy: The German economy and industry already struggles with digitalization. AI exacerbates this issue because it represents the next step of digitalization. Small- and medium-sized businesses in Germany, known as the Mittelstand, especially need support. This support could be, for example, through state-funded AI laboratories, in which companies can experiment with AI with little risk and at low costs. Mobilizing venture capital through public funds and providing better incentives for AI investments represent two more critical challenges. 6. Societal dimension of AI: The ethical and regulatory questions regarding AI need to be openly discussed and require input from many different stakeholders in German society. Here, we already see numerous initiatives and approaches, representing the topic’s arrival on the political agenda. However, more has to be done to make AI competencies and technologies more familiar within society. 7. A national AI strategy in an international context: Germany can only succeed in the international competition in the long-term as part of an EU-wide approach. Striving for cooperation with France offers the chance to push for a comprehensive European AI strategy. Germany, and Europe as a whole, have to become more conscious of their strategic interests in AI and act accordingly. A German AI strategy should focus on the ecosystem approach and propose concrete ideas and recommendations. The strategy also needs to address how we can identify and, if possible, measure how the strategy is being implemented, and how the AI ecosystem in Germany is developing. There are many important indicators that policy makers should consult in order to evaluate the effect of their actions: the attractiveness of German institutes and universities for leading international AI researchers, the number and quality of AI patents, achievements in publishing and visibility at the most important international AI conferences, venture capital investments, the founding of firms, or the number of companies with strong AI competencies and their growth. The good thing is that Germany does not have to start from scratch; numerous countries have already published national AI strategies in which many good ideas can be found. Now is the time for Germany to follow suit. Only then can Germany become a leader of AI development.

Full Text

Published Version
Open DOI Link

Get access to 115M+ research papers

Discover from 40M+ Open access, 2M+ Pre-prints, 9.5M Topics and 32K+ Journals.

Sign Up Now! It's FREE

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call