AI Strategy for Germany

by | Oct 25, 2018 | Artificial Intelligence, Business, Strategy

I was invited by the German government to participate in the development of the Artificial Intelligence (AI) strategy for Germany. More precisely to provide input on the topic how to apply AI research in industry and society.

Here are a few thoughts I have prepared for a workshop in Berlin last month on this topic.

Principles

There are a few principals which I believe are critical for a successfull implementation of the  AI strategy

  • Faster decisions and more flexible processes. Speed is of essence.
  • Use existing organisations if possible. Don’t reinvent the wheel.
  • Targeted investments – not the famous “Giesskannenprinzip” (where everyone gets a bit of the cake)
  • New investments are needed – not just reallocation of existing funds

1. Cutting-edge research

Europe and especially Germany is still well positioned in AI research, but not competitive anymore in the recruitment of top AI researchers (especially in the field of Machine Learning). This is a critical problem for the German (and European) economy. The development of artificial intelligence and its applications is still in its infancy. We need to invest in basic research that will drive the next wave of AI and train the top talents we need to develop the next generation of AI applications. To this end, we should investments in leading research institutions such as Max Planck Institutes, which can recruit top AI researchers from all over the world. This is the necessary prerequisite for a successful AI strategy in Germany.

2. Radical Innovations

Innovative companies looking for radically new products can use this objective to provide AI research with questions for new technical solutions that are still unthinkable today. The new German Agency for “Sprung Innovation” can play a key role here. New ideas must be developed and tested, which often means high risk and high investments. An important prerequisite: the new agency has to have the freedom to decide independently on its investments. Following an agenda predetermined by politics will not work . The resulting collaboration between cutting-edge AI research and innovative companies should be based on simple rules in terms of staff exchange and the use of data and IP.

3. AI Innovation Hubs

Medium-sized companies and start-ups are the engine of Germany’s (future) economic success. For many of these companies, the use of AI will not only be an important part of their strategy. It is also the basis for new AI-based solutions and processes, and thus for tomorrow’s market leaders.

For this to happen we need an AI Knowledge Transformation Initiative, in which AI Innovation Hubs are created at selected locations in Germany. These will specifically promote the transfer of AI knowledge to (medium-sized) companies and AI startups. An example of this already exists with the appliedai initiative.

These Innovation Hubs should receive funding from federal and state governments. In addition to imparting AI knowledge to companies, the establishment of Living Labs is an essential aspect. Here companies and startups can create prototypes together with AI experts. Thus new concepts are tested quickly and create the basis for future products. The cooperation of startups with their ideas and companies with their data and financial resources should be promoted in a targeted manner.

4. AI startups

In addition to the AI Innovation Hubs and existing Accelerator and Incubator programs, a new program modelled after the Creative Destruction Lab in Canada would be a sensible investment to provide AI startups with the necessary business know-how. Such a program could emerge in conjunction with the AI Innovation Hubs, involving top research organisations, investors, business angels and mentors. The goal is to support AI Startups with the potential to become global leaders.

In addition investors and venture capital firms should be encouraged to set up AI-focused funds in Germany, e.g. through capital gains tax relief.

5. Open Source

The open source model is well suited to quickly implement innovations in the field of AI. It allows new innovations to emerge more quickly through the rapid exchange of new ideas.

An open source incubator initiative in the field of AI should be started, with the goal to promote new AI startups based on open source principles.

6. Data

Data is the basis for every AI application. Availability and access to data are elementary prerequisites for a successful AI strategy.

Policy could, for example, provide tax incentives for relevant companies to make their data available. This should always be voluntary. They could provide funds for the purchase of hardware for large data centers and thus support data pools. Public data should be made available as far as legally possible. The legal framework (e.g. antitrust laws) has to support these new ways of data pooling.

Policy could also fund research programs aimed at a) generating data through simulation and b) finding new ways in which people can generate data.

7. Integration of society

The basis of a successful AI strategy in Germany (and in Europe) is a broad acceptance of the new technology among the population. The development of a basic understanding of artificial intelligence in society is therefore a critical success factor. This requires broad investment in education and training. Starting at school, continuing with education in companies and public institutions. To take away the fear and uncertainty many people has in regards to this new technology it will be critical to implement AI projects to showcase the benefits of AI solutions for society and individuals.