German Government released national AI strategy

By 20. November 2018News

German AI Strategy seeks more data exchange for research at national and EU level

In the race of becoming a global Artificial Intelligence (AI) powerhouse, Germany plans to invest three billion Euro in the application of the AI Strategy by 2025. Due to be officially presented during the Digital Summit in December this year, the strategy includes up to twelve spheres of activity where Germany wants to take the lead.

Significant investments in infrastructure

Heavy funding will be fed into new infrastructure that is designed to enhance research and help implement AI solutions in real-life scenarios. Same as for hardware “Made in Germany”, the country seeks to achieve the trademark “AI Made in Germany”, a label that would consolidate the country’s image as an innovation hub both in Europe and worldwide. To accomplish this, the government plans to build a cluster of 12 existing and to-be-established centers for research, development and application of AI solutions that will tempt specialists with attractive working conditions and salaries.

In addition to this, the federal government plans to establish a national research consortium that will act as a network of technology-, domain- and application-oriented establishments. Part of the network will be both the existing and upcoming competence centers, as well as academic research groups, research institutes, and data and IT centers. There are five national AI competence centers in Germany, with focus on Big Data, Machine Learning and IT Security. Four new machine learning competence centers are currently being started – in Berlin, Munich, Tubingen and Dortmund. They are designed to enhance research and deliver practical AI-based solutions.

Academia will mainly benefit from the strategy though the generation of 100 extra AI chairs that will ensure new specialists are being prepared for the upcoming AI-driven generation.

The creation of more real labs (testbeds), similar to Testfeld Autobahn A9, will also be supported by the government. The testbeds are experimental settings for pilot technology testing in a real-life environment. These will help developers test their technologies in settings similar to those from the real world without endangering people in cases of failure.

To ensure a smooth and efficient development and application of AI solutions, a German observatory for AI will be created. The institution shall monitor the impact of AI on the labor market, identify potential pitfalls and negative effects on society and combat any discriminatory practices by AI-run systems. In other words, the observatory is set to oversee the ethical issues of AI in Germany.

International cooperation: a “European AI Strategy”

At European level, Germany wants to focus its AI efforts in a joint project with France. A German-French “virtual center” is to be established in the near future. This will enhance the cooperation between the two countries by exchanging skilled workforce, organizing joint events, doctorates and summer schools. Moreover, bilateral AI clusters with concrete real-life application will be created for sectors like health, mobility, robotics and environment.

The German-French tandem will be urged to develop the nucleus of a general European cooperation in the area of data infrastructures and high-performance computers. The joint research and development of infrastructure is expected to lead into a “European AI Strategy”.

A European innovation cluster in the form of a EUREKA-Cluster and a European Data Space are further optimistic initiatives stipulated in the German AI Strategy. The Federal Government promises to support the EU Commission in the application of the initiatives, which is also set to strengthen the European data economy.

Legally-compliant data exchange

A priority of the AI Strategy is the secure data exchange and data availability for research purposes, given the need for quality data to train AI algorithms. To address this issue, the strategy suggests the establishment of legal “data partnerships” between companies and research institutions to share data. For this purpose, the visibility and support for the existing data platform International Data Spaces (IDS) will be increased.

An extra tool – the National Research Data Infrastructure (NFDI), which is currently under development – will ensure that German research sector will record, store and share the often decentralized, project-based and temporarily stored datasets with the German science, a move designed to contribute further to data standardization. A similar infrastructure – European Science Cloud (EOSC) – already exists at EU level.

Additionally, open-source data should be made available to researchers and developers through open-data platforms. Data protection is expected to be ensured though anonymization, and where possible through synthetization and pseudonymization. The health sector is one example where such privacy-enhancing technologies could be applied in order to protect individuals’ private data and informational self-determination. Furthermore, within the framework of the Medizininformatik Initiative, several data integration centers will be opened at University Hospitals throughout Germany, whose scope is to merge data from multiple sources, enhancing data protection and data sovereignty. This way the patients will have a better control over their data.

Dozens of German experts from the sector have been involved in the development of the AI Strategy. Two members of the advisory board of the Smart Data Forum – Prof. Dr. Volker Markl (TU Berlin) and Prof. Dr. Wolfgang Wahlster (DFKI) – have contributed to the conceptualization of the German AI Strategy.