Computer Science Embedded with Ethics

Computer Science Embedded with Ethics

News. Danish scientists in bioethics and philosophy are mapping ethical dilemmas in artificial intelligence (AI) science, developing a framework on responsible AI science and innovation practices, and building graduate studies in AI ethics.

“Data usage is like fish farms, agriculture and forestry. It is part of our fundamental existence. It is not either or. We cannot choose to leave the data-based society. As with fish farms, agriculture and forestry we have lots of challenges with data usage. And we may need to limit the usage in certain areas. That is why we need norms and ethics.”

Philosopher Sune Holm from University of Copenhagen’s Faculty of Science has begun a two year project on “Responsible AI in Science” (2019-2020) with his colleague, professor in bioethics, Peter Sandøe.

In recent years we have witnessed one scandal after another on data abuse, and one of the major consequences is diminishing trust from citizens and consumers.

“Our job is to help out with ensuring, that scientific use of big data and artificial intelligence in Denmark is done responsibly within our society,” says Sune Holm. ”It is also important to think about how big tech companies like Facebook and Google use and arguably misuse personal data, but our main focus is on Danish-based projects and the ethical challenges the scientists meet.”

One important task is to help develop an ethical code for artificial intelligence research at University of Copenhagen. “We aim to identify questions that researchers should think about in their work, not just a list of dos and don’ts”.

Part of this work will include mapping international research ethics such as Data Stewardship Accountability Elements (2018), Canadian Guidance on Accountability (2016), and EU Accountability Principles.

He will also collect cases such as the Challenge platform, where data from Statistics Denmark and digitalised human tissues are analysed by machine learning to better understand ageing. Or this project, Analyzing Partitioned FAIR Health Data Responsibly, from Maastricht University, where health data and data form Statistics Netherlands is used in machine learning to better understand the relationship between the development of diabetes and socio-economic factors such as lifestyle and health care utilization. Both projects are striving to do it in a privacy-enhancing manner

Finally, the project will also develop teaching in ethical considerations when analyzing big data for science students at the University of Copenhagen. The aim is to give future generations of data scientists a solid understanding of the complex issues of their work and hence contribute to the development sensible technology and responsible use of data in Denmark.

“We need to establish a conceptual framework for discussions on handling data and point to best practices. Why do we go there? Why and how did we reach that conclusion. Did we do this in an ethical way? How should we balance possible societal gains with protections of people’s privacy and integrity?,” says Sune Holm.

No Comments

Post A Comment