How to develop an ethical AI

Students share their experiences from the EELISA “AI, Ethics & Human Rights“ summer school in Madrid

An artificial intelligence that can use medical data to predict how likely a person is to become seriously ill in the next few years. With this knowledge, the so-called digital twin can make recommendations – to change diet or start a certain sport – and thus prevent the disease from breaking out in the first place.  But what happens to the data fed into the AI? Can health insurance companies get access to it? How do patients deal with the predictions? What are the negative effects of using this technology? “So far, I have been fascinated above all by how artificial intelligence works and what positive effects it can bring to medicine, for example,” explains Victoria Müller, who is studying medical technology in her master’s degree at FAU. “But in this context, it is so important to think about the impact artificial intelligence can have on humanity already during the development of AI.”  At the five-day summer school “Artificial Intelligence, Ethics & Human Rights” at UPM in Madrid, Victoria had the opportunity to do so.

(Photo: Serena Bischoff)

“We were students from Italy, Spain, Turkey, Peru, Mexico and Germany, all with different scientific backgrounds. This interdisciplinary and international exchange was so exciting. In some cases we had very different opinions and really discussed a lot,” explains Zekiye Bersu Sezer, who studies electrical and electronics engineering and also participated in the summer school in Madrid. To create a common knowledge base, the summer school started with an introduction to both ethical analysis techniques and the scientific background of Artificial Intelligence.

The students then worked on different projects in small groups: Under what circumstances can artificial intelligence be used for diagnosis in imaging procedures such as MRI or CT? Should the police use artificial intelligence to predict future crimes from past ones? Using a guideline, the students discussed the various implications: for example, which people might be disadvantaged or who is legally responsible in case of wrong decisions. At the end, they all presented their results: Frameworks for the ethical use of AI in the respective project.

(Photo: EELISA)

“In the discussions, I found that while many use the keyword AI, they often don’t know how the different types differ,” Victoria explains. “That’s why I decided to also explain the basic methods of the different AI systems in my video.” In order to receive credits for the summer school, the FAU students produced short videos for FAU’s Digital Ethics Lab. “I learned so much during the summer school, especially how to think through topics critically and systematically. With the videos, we can pass on this knowledge. Understanding how AI works and the challenges of using it is also interesting for the general public – after all, we are confronted with AI everywhere in our everyday lives,” explains Zekiye, who discusses the use of AI in student assessment in her video.

“But not only the lectures and the project work were exciting, the atmosphere in the group was also great. After the course, the students from Madrid showed us around the city, and we went out for dinner and had a drink together,” says Victoria. She is still in contact with some of the participants.  “It was really a great experience that I was able to have thanks to the EELISA sponsorship. I made new friends there, but also got to know fellow students from higher semesters whom I can ask for advice and got in touch with great keynote speakers and researchers,” explains Zekiye.

Interested in a summer school or workshop abroad?

Students can regularly apply for activities organized by EELISA partners in Paris, Pisa, Erlangen, Madrid, Bucharest, Budapest and Istanbul. EELISA offers financial support for travel and accommodation. The current activities are available on the EELISA website or in the communities on the EELISA Community Platform.