The UK intends to authorise self-driving cars from 2021. In Germany, politicians and businesses are still rather sceptical about AI and about digitisation in general. What can Germany learn from other countries?
First of all, it is essential not to be sceptical but rather to be a proactive about it at national and international level. In the face of global competition, many countries, including the UK and Finland, have launched national initiatives to advance the development of AI. Like other countries, Germany needs to engage with this topic proactively and keep pace with the front-runners. We are well-positioned, but lack a coordinated agenda that systematically addresses issues like capacity building and the skills shortage. Also, AI is a cross-cutting sector. So Germany should be doing even more to promote cross-sector innovation. Up to now, information has been shared largely within sectors. Legislators need to actively deal with the issue and advance it. Overregulation is an impediment to innovation.
But much has also happened in Germany: Karlsruhe is home to Germany's oldest informatics faculty, the Karlsruhe Institute of Technology (KIT), which is very user-oriented, especially in the field of robotics, and two Fraunhofer Institutes working in this field are also based in the city.
Furthermore, initiatives like appliedAI by UnternehmerTUM, the Center for Innovation and Business Creation at the Technical University of Munich, bring together companies, start-ups and research.
German companies, including VW, recently even warned against artificial intelligence. What is your take on this?
For nearly twenty years now, I have been working on innovation, and one of the fundamental principles that I have learned during this period is openness, i.e. a positive attitude to change and new ideas.
However, people here seem to be rather afraid of change. In the near future, AI will revolutionise not only manufacturing, but also the organisation of entire companies – due to AI, hierarchies may become blurred and routine processes and decisions can be supported or even taken over by AI. People are freed up from work in these fields to fulfil more creative tasks.
AI of course also brings challenges – like the decision whether a self-driving car should give way to an obstacle and put its own passengers in danger, or should keep going, potentially injuring passers-by. And who will be held liable when there’s an accident? The entire ecosystem needs to work together to find answers to these questions.
The response to scepticism and doubts has to be solutions-oriented action. In the field of artificial intelligence, there are certainly many issues that need to be addressed (including ethical principles, privacy, control mechanisms and responsibility).
What sectors do you expect to be revolutionised in the coming years, and what is the role of AI in this respect?
Digital transformation will fundamentally change all sectors. Every process step that can be automated through data provides an opportunity to use artificial intelligence.
Some industries, the media sector for instance, are already very advanced in this respect, while others, including finance and manufactuing, still have a long way to go. Digital success will always involve effective and sustainable data and AI solutions. On the one hand, AI helps to make key business processes like distribution, marketing, finances and interfaces more efficient, e.g. a company's chatbots and virtual assistants. Even more importantly, AI is creating the basis for new smart services like autonomous driving and flying, predictive maintenance, fraud detection, text and video analysis and interpretation (e.g. in the medical sector).
And where will it be difficult to use AI?
There is no one specific industry that is less suited to AI. But industries and fields involving complex, multidimensional decision trees make it much more difficult to use AI. In general, there are several AI levels (see diagram). In some areas, levels 2-3 can already have significant effects – at these levels, we often talk about augmented intelligence – while in other areas new use cases can only be realised at levels 4-5 – e.g. fully autonomous driving.
Medicine is an interesting field – computer vision, for instance, can provide valuable support for doctors when it comes to diagnosing diseases by identifying the relevant symptoms on x-ray images. For this purpose, however, the relevant algorithms need to be developed and continually improved step by step and in close cooperation with experts. As I mentioned above, it is important to understand that artificial intelligence is an iterative process rather than a solution. Wrong interpretations in the medical sector can have far-reaching consequences, so the development of a fully autonomous system (e.g. for surgery) is rather unlikely.
To what extent can the Digital Hub Initiative and the digital hubs contribute to advancing digitisation in Germany and the use of artificial intelligence?
The expansion of cross-sector AI ecosystems is a decisive element for the future development of AI. Close cooperation between research, start-ups and established companies supports talented people, helps develop innovative use cases and thus lays the foundations for future competitiveness. The way the digital hubs are structured and located is perfect for the various industries. Success stories from the Artificial Intelligence Hub in Karlsruhe are being adapted and applied in the other digital hubs. The digital hubs can make a significant contribution by actively promoting and supporting the establishment and expansion of AI within the hubs. Furthermore, the hubs can really help us develop a common position and an action plan for government and decision-makers.
Germany is facing challenges in a global environment, and it will need a smart and coordinated plan if it is to compete.
DAIN Studios has offices in Helsinki, Berlin and Munich. It helps companies develop their AI and data strategies and implement AI services and solutions.
