| News | Interview | “It’s always impressive to realise what humans can do – and machine learning cannot (yet) do.”

“It’s always impressive to realise what humans can do – and machine learning cannot (yet) do.”

Ute Schmid, Professor of Applied Computer Science, esp. Cognitive Systems at the University of Bamberg, has been a member of the bidt Board of Directors since February 2020.

You are a psychologist and a computer scientist. To what extent does this interdisciplinary combination shape your research?

Very clearly on some topics. It makes sense to have a basic understanding of how human learning works when developing machine learning – sometimes to bow humbly to what humans can do and machine learning cannot.

An empirical view of human learning processes is also valuable for getting algorithmic inspiration, as an impetus for developing machine learning.

A basic understanding of how human learning and thinking processes work is also essential for developing functioning interfaces for human-AI partnerships, i.e. the interaction of humans and artificial intelligence. Purely autonomous applications of machine learning are not even possible in many areas, especially not in safety-critical areas where transparency and traceability are indispensable, such as in medicine.

You are also committed to teaching computer science to children. What motivates you?

When I came to the University of Bamberg in 2004, at that time as the only woman in the faculty, only about ten percent of the students were female. That was the trigger for me to get involved in giving girls the chance to discover and develop inclinations and talents in the field of computer science. Many studies suggest that it is crucial to start introducing children to STEM subjects, including computer science, at an early age.

At the moment, I find it particularly exciting how the topic of artificial intelligence can be taught in an age-appropriate way. For children of primary school age, I have designed an educational game that makes the basic concepts of machine learning vivid and comprehensible. For young people aged 12 and over – and perhaps also for adults without a corresponding background – I have written the book “KI selber programmieren” (Programming AI Yourself) with two colleagues, which introduces basic areas such as inference in semantic networks, machine learning, games, language and emotion in a very concrete way.

This can also make it understandable that pure pattern recognition is not the same as actual understanding, even if it sometimes seems with a chatbot that it understands what is being said.

How do you assess bidt’s mission to seek dialogue with society about the effects of digitalisation?

I think dialogue with society is extremely important. I believe that the development that artificial intelligence and machine learning are taking will affect all areas of society’s life. It will determine how we live, learn, work, are cared for. That is why it is so important that science and politics are in dialogue with citizens about how we want to live as a society with AI.

Do you have any plans for your involvement in bidt?

I am very much looking forward to the dialogue with the top-class interdisciplinary board of directors at bidt and to contributing the topic of artificial intelligence here. I am sure that I will also take away a lot myself.