Interpretation of German Artificial Intelligence Strategy

The German Federal Government’s Artificial Intelligence Strategy Report covers the development and wide application of AI, and the possible changes in political, economic, cultural, security, legal, ethical, and international cooperation, and identifies “weak” AI and “ Strong “AI” direction, clear that the government should include AI in all policy areas and propose action measures.

On November 15, 2018, the German Federal Government officially issued the “German Federal Government Artificial Intelligence Strategy Report” (hereinafter referred to as the “Report”). The “Report” proposes that artificial intelligence (hereinafter referred to as AI) has entered a new stage of maturity, to make “AI Made in Germany” a globally recognized product identity; to implement the “AI” with the concept of ethics and morality as a whole. Made in Europe” trademark. The “Report” covers the development and wide application of AI, and the possible changes in politics, economy, culture, security, law, ethics, international cooperation, etc., and proposes action measures. The “Report” is also part of the German digital implementation strategy, and will be further revised and improved by 2020 to cope with the rapid changes in AI.

Determine the recent AI direction

The German side did not reach a consensus on the definition of AI, but determined the two directions of “weak” AI and “strong” AI.

“Strong” AI means that the AI ​​system has intelligence similar to or even beyond humans. The “weak” AI focuses on solving specific application problems based on mathematical and computer science methods, and the system thus developed is self-optimizing. In addition, the researchers modeled and formalized all aspects of human intelligence and built systems for simulating and supporting human thought.

In order to seize the opportunity, the German federal government’s strategic orientation is to use artificial intelligence to solve the application problem, and proposes to locate the “weak” AI direction in the near future:

Inference system, machine proof. A system that derives formal statements from logical expressions to prove the correctness of hardware and software.

Based on knowledge systems. A method of modeling and bringing together knowledge, used to simulate human expertise and expert-supported software (expert systems), partly related to psychology and cognitive science.

Pattern analysis and pattern recognition. General inductive analysis methods, especially machine learning.

robot technology. The automatic control of the robot system is also an autonomous system.

Intelligent multi-mode human-computer interaction. Analyze and “understand” language (in conjunction with linguistics), images, gestures, and other forms of human interaction.

Clarify development goals and ideas

The German government proposes three major goals: to create Germany and Europe as AI locations, to ensure Germany’s future competitiveness; to achieve responsible, common well-oriented AI development and application; to embed AI in moral, legal, cultural and institutional society.

It is clear that the government should incorporate AI into all policy areas and implement a people-oriented, socially beneficial-oriented application at the economic and social level based on democratic order. The main idea:

Utilizing the advantages of existing scientific research and economic structure, in the leading position of technology in key industries such as Industry 4.0, combined with the potential of AI, through strong alliances, improve existing efficiency or enter new business models to realize future value.

Under the premise of respecting the self-determination of citizen information, the AI ​​technology system is embedded in complex products, services or business models, and trust relationships based on active participation, transparent procedures and traceability of citizens are established.

AI development offers great opportunities for traditionally strong medium-sized manufacturing, but it is a challenge for SMEs; AI has great potential for public management and certain government tasks.

Strengthen the discussion of the legal framework for AI; raise the awareness of the whole society and strengthen the factual and evidence base needed for public debate; the national strategy should contribute to “AI made in Germany”, contribute to national and social well-being, and be easily adopted.

Strengthening cooperation in Europe and internationally to ensure the success of the AI ​​economy, especially the unified ethical standards; the EU General Data Protection Regulations (DSGVO) is an important basis for European cooperation. In addition to realizing technological strength and utilizing market forces, it is also necessary to actively promote Its values; participation in the development of international rules and the promulgation of standards within the EU.

In order to effectively implement the high-tech strategy 2025, AI is the focus of the German innovation location and has become the mission of the government.

Launched AI action measures

From 2018 to 2025, the government invested about 3 billion euros in AI research and development, reaching 3.5% of research and development targets, of which the budget for 2019 is 500 million euros. The leverage effect will at least double the government’s disposable funds. In addition, beyond government powers such as education and training funding, special funding from national and social partners is required.

The Report proposes 12 actions: strengthening research in Germany and Europe, becoming a leader in innovation; innovation competition and EU innovation cluster; transformation of results, strengthening the role of SMEs; Fourth, awakening entrepreneurial vitality and helping its success; achieving employment and labor Structural adjustment of the market; strengthening vocational education, cultivating professional talents and experts; using AI for administrative management to provide management capabilities; ensuring data availability, reducing application barriers; adapting to regulatory frameworks; determining standards (who sets standards, who has market decisions) Power); cooperation at the domestic and international levels; strengthening social dialogue and further development of the political framework. Among them, the action measures related to science, education and industry mainly include:

Each year, we will establish an association with at least 1,000 companies through the “AI Trainer”; set up at least 12 centers and application clusters across the country; set up AI majors in universities, and launch science young scholars training and teaching programs, at least 100 new Professor; use AI basic knowledge as a vocational education and training content; develop AI technical worker strategy; promote AI and core technology, such as in-process technology and production technology; develop and apply AI in agriculture and food chain.

It is proposed that modern information communication systems constitute the central nervous system of the digital economy and society. In order to ensure network efficiency and reliability, and to efficiently and sustainably utilize existing resources, it is necessary to develop new learning and adaptive communication systems. In order to detect anomalies in networks and information communication and production systems, it is necessary to analyze massive amounts of data in a short period of time, and new (partial) automation programs based on AI must be researched and developed. Application of AI, taking into account the legitimate rights and interests of patients in the health field, data protection development for data generated by distributed data sources; promoting development in the field of nursing; strengthening the research and development of aerospace technology. For the AI’s existing research funding procedures, the government must set up a review period, and from the early stage of research and development, it is necessary to supervise whether it meets the professional requirements of the regulations, especially strict regulatory requirements.

Emphasis is placed on industrial AI applications that provide tools for efficient processing and processing of industrial process data. AI is a core element and key to intelligent monitoring, control and regulation processes that take Industry 4.0 to the next level. Through innovative business models, digital services and AI applications can be provided for industrial products; AI results are rapidly transformed in the industry and play an important role in the widespread application of the economy. SMEs should be more involved in areas such as AI technology, computing power and cloud platforms; open standards can support the widespread use of AI. In addition, machine-generated data and domain-specific data in industrial and B2B environments are particularly important, and data protection regulations must be observed.

It is proposed that the application of AI in all modes of transportation can make “people and things” move safer, more environmentally friendly, more efficient and cheaper. If a dangerous situation can be identified earlier, a more reliable response than humans can significantly reduce the number of accidents; future traffic flows can be directed, no more congestion or delays; no longer worry about parking spaces; logistics can optimize capacity , eliminate unnecessary empty loads, etc. The AI ​​system will play a key role in achieving reliability in the autonomous driving and automotive sectors. The generation/recording, management and evaluation of driving and sensing data promotes close cooperation between the automotive industry and the reliable application of the autonomous AI system. Work with data protection regulators and industry associations to develop privacy-compliant AI system development and application guidelines, present best-of-breed application examples; enhance and expand industrial data exchange and interoperability research; drive data format and interface standards development, and Cooperation in Europe, etc.

In summary, it can be seen that although the “Report” was published more than one year later in the United Kingdom, the United States, China and Japan, the “Report” was formed on the basis of drawing on the opinions and suggestions of the recent development of AI in various countries around the world. The proposed goals, ideas and action measures are more comprehensive, more pragmatic and more operational, and put forward specific requirements for the realization of “AI Made in Germany”. These are worth learning and learning from, and we have the initiative in the future competition.