When the Cabinet meets on Wednesday and Thursday for the closed-door meeting in Potsdam, nothing less than Germany’s future viability is on the agenda. The Artificial Intelligence strategy is to be decided. The paper in the Tagesspiegel, with its 80 pages, is almost a small coalition agreement. Although some concrete measures can be found in it – in some points the strategy shows weaknesses.
It wants to invest three billion euros by 2025 in order to make Germany “one of the world’s leading locations for AI”. That may sound like a lot, but can be seen in an international comparison: France is 1.5 billion euros in AI by 2023, China wants to create by 2030 a government-sponsored 150-billion-dollar industry.
For the time being, only € 50 million is earmarked for AI research in 2019, € 450 million for subsequent years. 116 million euros are added to the Agency for Jump Innovations – that even more money is added is not excluded. The government is already demonstrating its ambition with the strategy to apply for additional funds. In any case, it assumes that its funding will also trigger investment in industry, science and in the federal states, so that it will “at least double the funds made available by the federal government”.
AI research itself is already excellent in Germany . To keep it that way, at least twelve centers should join together to form a network. From the originally announced German-French research center in physical form, however, only a “virtual center” is left in the end.
In order to hold top researchers and attract them from abroad, the government wants to create an attractive “overall package”. So there should be a dual-career model, which makes it easier for them to switch between science and industry. This addresses one of the big sticking points: Excellent research has so far rarely and too slowly found its way into marketable products and services by German companies – just like the MP3 player, whose technology was developed in Germany, but the money is companies from Asia.
But no concrete solution is found in the strategy of how KI top researchers can be better paid in the future. This would require an adjustment of the civil service collective agreement (TVöD), but so far there have been no allegations of any talks between Verdi and the civil servants’ union with the Ministry of Economy. Here the government would have to reload.
Researchers not only need good working conditions, but above all data. Therefore, the government wants to promote research for pseudonymization and anonymization. If convincing models are developed here, more citizens should also be prepared to make their data available for research purposes. In addition to real data, the government also wants to push ahead, however, that data can be generated more easily in the future synthetically. Companies themselves are to enter into “data partnerships” with the involvement of the antitrust authorities.
This should then also benefit the middle class, who should get in the matter of AI tuition from so-called “AI trainers”. At least 20 such trainers should advise and assist the small and medium-sized companies with the implementation of AI. The effect of such measures is so far difficult to measure. This should change now: With an AI monitoring, the government wants to determine the penetration of AI applications.
A German AI observatory is also to be founded, which should in particular observe how AI affects the world of work. In the future, it could also act as a supervisory body and, for example, check application and personnel planning software for discrimination. There was no agreement between the ministries on a separate employment data protection law. Here is probably 2019 refilled.
In any case, at the beginning of 2020, one year before the end of the current legislative period, the government intends to review how the strategy works and whether and how it has to be adapted to the new developments – so that “KI – made in Germany” becomes a seal of quality.'[‘