Before the summer holidays, three online seminars in English on responsible artificial intelligence are planned. Aasa Feragen and her project partners in 'Bias and Fairness in Medicine’ hope to connect the AI researchers across Denmark and universities.

Webinar series: Researchers are focusing on responsible AI

Monday 03 May 21

Contact

Aasa Feragen
Professor
DTU Compute
+45 26 22 04 98

Contact

Lars Kai Hansen
Professor, head of section
DTU Compute
+45 45 25 38 89

Join Responsible AI Seminar (English)

  • Friday May 7 2021 15:00-16:00 Lars Kai Hansen (DTU): Values in AI
  • Friday May 28 2021 13:00-14:00 Veronika Cheplygina (ITU): How I failed machine learning in medical imaging - shortcomings and recommendations
  • Friday June 18 2021, 13:00-14:00 Jens Christian Krarup Bjerring, (AU): Black-box decision making in medicine: some thoughts and questions

Join on Zoom.

Learn more about the seminar series

 

 

About 'Bias and Fairness in Medicine'

  • The three-year project ‘Bias and Fairness in Medicine’ researches fairness in artificial intelligence in medical applications where AI is used in the diagnosis of diseases.
  • The researchers will test a hypothesis that distortions in the data could mean that population groups are underdiagnosed because the health care system and the AI methods use most data from a particular group. The goal is to develop fair algorithms so everyone gets fair treatment in healthcare.
  • The project is a collaboration between DTU Compute, the University of Copenhagen, Rigshospitalet (Denmark´s leading hospital for patients needing highly specialized treatment), and the technical university ETH in Switzerland.
  • The project is supported by the Danish Independent Research Foundation.
  • Learn more about the AI seminar series and the project.

The critique of artificial intelligence dominates the debate. A series of seminars across the Danish universities will try to nuance the picture of AI to remember the technique also holds many good opportunities

Recently, the media has brought more and more stories about how apps and websites, using artificial intelligence, collect, analyze, and exploit our behavior to hold our attention and try to influence our attitudes.

But artificial intelligence also holds many possibilities. Last week, for example, the European Commission presented a package of rules and measures 'to turn Europe into the global hub for trustworthy Artificial Intelligence (AI). The EU will strengthen investment, innovation, and the spread of AI in a way that guarantees security and fundamental rights for EU citizens and businesses. We must have confidence in artificial intelligence.

With a new webinar series on responsible AI, the initiator, the people in the project ‘Bias and Fairness in Medicine’, hopes to make Denmark a frontrunner for responsible artificial intelligence. The plan is for researchers across Danish universities to debate various facets and issues within AI, says Project Manager and Professor of medical image analysis at DTU Compute Aasa Feragen:

“Right now the challenges are getting a lot of attention and it is important to be open about them. But we must remember that AI also offers new opportunities. The techniques we use for facial recognition are the same we use to recognize tumors on scans, and we would like to use AI for that.”

"Now the EU wants to make laws and regulations that will regulate these things. It is needed, but it is important that we can still develop the good things as well as possible so that we can protect ourselves from e.g. diseases, ”says Aasa Feragen.

Starts with the value debate

"The techniques we use for facial recognition are the same we use to recognize tumors on scans, and we would like to use AI for that."
Project Manager and Professor of medical image analysis at DTU Compute Aasa Feragen

Ultimately, responsible AI is about what values we think are important in our society. Privacy - versus what we can achieve with a lot of data. What options are there for technology versus what damage it can cause.

In the first lecture in the seminar series on Friday 7 May Professor and Section Leader for Cognitive Systems at DTU Compute Lars Kai Hansen will talk about values and AI:

"The technical aspects of ethics, such as bias, transparency, and privacy, have gained much attention in the AI community. But there is an increasing realization that we need to discuss the more fundamental issues of "power" and "values"."

In his lecture, he will discuss how AI is shifting power and the possibilities for understanding the value sets of AI systems, and the opportunities we have to work with values in an AI context. The value debate has become extra important due to social media, he believes.

"AI systems are not value-neutral. Social media play an increasing role in public discourse and the values they promote may have massive societal impact. Are Facebook's values e.g. aligned with yours?" Lars Kai Hansen asks.

Before the summer holidays, three online seminars in English on responsible artificial intelligence are planned. Aasa Feragen and her project partners in 'Bias and Fairness in Medicine’ hope to connect the AI researchers across Denmark and universities.

The idea is that the initiative will continue after the summer holidays with a new series of AI seminars online, where the universities take turns addressing relevant topics.