Artificial intelligence is not a replica of human intelligence

Professor Dr Hanno Gottschalk on artificial intelligence research at the University of Wuppertal

The mathematician and physicist thinks little of Sophia, a humanoid robot developed by the Hong Kong company Hanson Robotics, which stands out from previous robot variants due to its particularly human appearance and behaviour, because it is capable of visual data processing and facial recognition, imitates human gestures and facial expressions and is also able to answer certain questions and hold simple conversations on predefined topics. "I don't need it," is his short answer, although artificial intelligence itself is a particular focus of his work.

The scientist, who specialises in applied mathematics, came across the topic of artificial intelligence (AI) through his students. "I teach statistics and we have developed a seminar in which we go to small and medium-sized companies in the region and offer data analyses for them," he explains. Specially assembled teams go to the companies, are given a specific task and then present their results. "Over time, more and more technologically advanced companies have become involved in this format" and Gottschalk initiated a network. Together with colleagues from the Faculty of Electrical Engineering, the Velbert/Heiligenhaus Campus (CVH) of Bochum University of Applied Sciences and data scientists in regional companies, he founded the Bergisch Innovation Platform for Artificial Intelligence (BIT). With joint research, student training and close networking between universities and first-class technological companies and users, the declared aim is to make the potential of AI and data analytics accessible to the region.

The digitalisation strategy of the University of Wuppertal

In order to advance the digitalisation strategy, the rectorate of the University of Wuppertal has set up the development and implementation of a comprehensive IT development plan for the University of Wuppertal as a strategic project. The guiding concept of the digitalisation strategy used for this is intended to express the fact that it deals comprehensively with the objectives and forms of the use of IT tools and services in all functional areas of the University of Wuppertal. To this end, three guidelines for the areas of research, teaching and administration/management were adopted for the digitalisation strategy. There is a new computer science bachelor's and master's degree course with a focus on data science in the schools of mathematics and natural sciences and electrical engineering, media technology and information technology, six new professorships and the newly founded Interdisciplinary Centre for Machine Learning and Data Analytics, of which Gottschalk is the founding spokesperson together with the Chair of Electrical Engineering and Theoretical Communications Engineering, Prof. Dr-Ing. Anton Kummert.

AI is not a replica of human intelligence

Many people are sceptical about this new development. The mathematician says: "As with any technology, there are sensible, less sensible and extremely problematic applications of AI. Even if certain impulses have come from neuroinformatics, AI is not a replica of human intelligence. I think a lot of the fear comes from the claimed similarity to human intelligence, which is more of a marketing gimmick than a reality."
The Bückeburg native does not see the danger suggested in many apocalyptic films that humans will one day lose the upper hand over these machines, which may take on a life of their own. "If we think in terms of films again, world domination always requires a very strong will. And what AI doesn't have at the moment is anything like intention and will. It still obeys the goals set by humans. The question is, which humans set the goals? With what intentions and with what will?" What is much more interesting to him about this new technology is the question of "who owns it and who owns the opportunities that result from it? This raises much bigger social questions than these visions of robots taking over the world."

The BIT begins its work with 2.7 million euros in funding

The new Bergisch Innovation Platform for Artificial Intelligence is funded by the state and the EU and began its work after the kick-off event in March. "We have three groups of participants. These are Bochum University of Applied Sciences, Velbert Heiligenhaus Campus, the associated companies, which are also investing 700,000 euros of their own funds, and the University of Wuppertal, which is in charge and receives the largest amount of funding." Together with the advisory board, Gottschalk is developing formats for cooperation between regional industry and research using the example of AI. He calls this enabling technology, i.e. technologies that, in combination with other technologies, generate significant leaps in performance and capabilities for the user and thus enable innovation and transformation in medium-sized companies. And according to the scientist, virtually all areas of society benefit from these developments. AI is already being used in all areas of technology and production. On the one hand, old technologies are being carefully scrutinised with "new methods of predictive maintenance or quality control", while on the other, "technologies that have not yet been invented, such as the process towards autonomous vehicles" are being further researched. Gottschalk explains very clearly: "There are things where society has to decide whether it wants them or not!" As examples, he cites the intensive development of care robots in Japan or the prediction of academic success for students in the education sector.

Cooperation between research and companies

As the university itself does not have any data, the scientists need data material from the companies. In this way, collaborations are entered into on a basis of trust, taking data protection into account. Many companies also offer co-operation in the preparation of bachelor theses or are looking for interns. This allows students to familiarise themselves with the processes of IT companies on site. Another synergy effect is the exchange with the company's own data scientists. "Last but not least," emphasises Gottschalk, "the question of spin-offs is relevant. Many of the data scientists we are now training here in research will also have the opportunity to found companies. And we are particularly pleased that we have many entrepreneurs on board who can advise and coach them accordingly."

Safety of autonomous driving

The main topics are wide-ranging. From use in pure production to automotive and mobility applications, where Gottschalk's working group is based. "We are working on the safety of autonomous driving! We are primarily concerned with the faults of AI and the limits of its usability. If we can't recognise these more precisely, it's problematic to use AI in safety-relevant areas." Under defined, controlled circumstances, the 52-year-old can already imagine that in five to ten years' time he might be able to drive from Bochum to Wuppertal on the motorway, "and I'll take a little nap and then arrive at the students even fresher," he predicts with a smile.
He is particularly pleased about the collaboration with the companies from the Bergisch Innovation Platform, which provide the AI service, the Institut für Qualitäts- und Zuverlässigkeitsmanagement GmbH (IQZ), which takes care of the validation of AI ideas, and the start-up company Instana from Solingen, which he hopes will provide good impetus towards founding a company. "These are all great companies. It's a dream to work with them and I'm really keen to do so."
"The real question is not whether machines can think, but whether people do," says US psychologist B.F. Skinner. Hanno Gottschalk and the scientists at the Bergisch Innovation Platform for Artificial Intelligence can.

Uwe Blass (interview from 10 April 2019)

Hanno Gottschalk completed his habilitation at the Rheinische Friedrich-Wilhelms-Universität Bonn in 2003 and then worked at Siemens AG in Mühlheim an der Ruhr, among others. He has been teaching and researching in the Stochastics working group at the school of mathematics and natural sciences since 2011.