Whether it's a voice assistant, a vacuum cleaner robot or facial recognition on a smartphone - AI has already arrived in all areas of life. Artificial intelligence has advanced to become a central driver of digitalisation. Compared to previous social and technological changes, the speed of this process is unprecedented.
DM: Dr Hecker, in which areas does AI make sense?
Dr Dirk Hecker: The Covid crisis has shown us how important digital infrastructures are to keep socially relevant processes running, e.g. in hospitals, educational institutions or public administration. We need to make improvements here, and I see a lot of potential for AI systems. For example, in the processing of forms in public authorities, but also in business, we are still very much characterised by manual activity. Intelligent document analysis can streamline processes here so that workers can perform more relevant activities, for example. AI not only helps to group documents with a high degree of quality, to extract information in a targeted manner, but also to translate them.
One of the best translation programmes currently on the market is DeepL from Cologne. You will be surprised how good the quality has become. Another example is the automatic subtitling of television programmes for the hearing impaired. Fraunhofer IAIS has developed an AI application for this, which is used, for example, for live subtitling in the Saxon parliament. The broadcaster ARD also uses our AI systems for speech recognition to search audiovisual content for quotes, for example. In general, wherever large amounts of data have been collected or will be collected in the future, AI applications can provide support. There are hardly any limits to the variety of application areas, the potential is immense. In addition to continuous improvements, AI also has the potential for real disruption that completely overturns established markets.
DM: The Fraunhofer-Gesellschaft develops key technologies for AI and its applications at many institutes. Who are your clients?
Dr Dirk Hecker: The Fraunhofer-Gesellschaft is Europe’s leading organisation for application-oriented research. As part of this, we at Fraunhofer IAIS work on public research projects, both national and EU-wide, such as the current AI4EU project, which is about the development of a European AI ecosystem. On the other hand, we work for clients from industry. Companies turn to us to benefit from our experience and to close their own competence gaps: from technical consulting, know-how building among employees to the development of AI solutions. When it comes to implementing AI solutions, American companies are very present globally, if not dominant. Unfortunately, the race for the big internet platforms in the B2C sector is already almost lost.
In many sectors, however, AI is still in its infancy, especially in the industrial sectors, where Germany has its strengths. We need to move forward here and reflect on our strengths as a high-tech country. To do this, we need AI ecosystems made up of industrial companies, universities, start-ups and applied research institutions that create an efficient transfer.
With a similar eagerness as German engineering was once established as a global brand, the new territory of data-driven services and business models can now be actively shaped. The vision: precise high-performance sensor technology and powerful machines are combined with intelligent AI systems "made in Germany" and exported worldwide. We have the opportunity to strengthen existing markets and build new ones. To do this, we must master the key AI technology, train skilled workers, offer attractive opportunities for AI researchers and also always be courageous in rethinking existing processes.
DM: On your website it says: "A deficit of one's own competencies leads to dependencies on actors who, for example, have lower standards of data protection, security and algorithmic traceability" What exactly is meant by this?
Dr Dirk Hecker: There is no guaranteed success recipe for innovation, but there are drivers and technological accelerators of progress. AI know-how is one such accelerator that helps determine competitiveness in the national and international context. If we leave this competence to others, then we will have to live with the fact that decisive technology modules will no longer be our responsibility and sovereignty in the future. This must be seen in particular against the backdrop of data protection and ethical values for which Europe stands. However, to guarantee digital sovereignty and market-economy independence in Europe, politics and business in this country must invest even more in AI know-how. There definitely is pressure to act so that Germany can continue to shape the digital world market in the future.
To take up the analogy again: Machines from Germany stand for a special quality. That's also how we want to build AI systems. They are transparent and trustworthy in use, run robustly, respect safety standards, do not discriminate and preserve data protection. This strengthens trust in AI technologies and represents another German and European seal of quality.
With the Fraunhofer-Gesellschaft and other scientific institutes, we have already built up a powerful research landscape in the field of AI in this country, and companies can dock here at any time. In terms of standardisation and trustworthy design of AI, we are also making progress in Europe. At Fraunhofer IAIS, we recently presented the first AI audit catalogue as a structured guideline for the design and quality assurance of AI applications. I hope that AI systems from Germany will become the same export hit as the machines of our engineers, which are distinguished by their special quality.
Interview Marie Wildermann