Fika-to-fika workshop: Human-friendly AI

When: 1 March at 9.30 to 15.30 CET

Where: Eden Hörsal (prel), Paradisgatan 5H, Lund, Sweden

Registration:  Participation is free of charge, but you need to sign up at ai.lu.se.

Description of the day

While the development and proliferation of Artificial Intelligence (AI) technologies in all spheres of society has been marked by great optimism, there is a growing concern among both scholars, political actors and the public that AI can disrupt or even harm the rights of individuals and our democratic societal structures. For example, several research projects have shown that AI-based applications may enforce gender and racial bias, political actors have (mis)used AI technologies to micro-target and influence voters in elections, and machine-learning algorithms have been accused of facilitating the creation and dissemination of disinformation. Similar concerns have been raised about the societal disruptions and job displacement caused by AI-enabled forms of automation and robotics.

These concerns have informed a growing scholarly interest in so-called human-friendly AI technologies. This growing interest is evidenced for example by a forthcoming EU HORIZON call on ‘human- friendly deployment of artificial intelligence and related technologies’, the aim of which is to promote research on both the societal risks associated with AI technologies and the development of AI-systems that are trust-worthy and protect the rights of individuals.

But the growing interest in human-friendly AI technologies also raise new questions. For what are the reasons that we even need human-friendly AI in the first place? Can a robot ever be “friendly”? And what challenges and societal risks does the development and dissemination of such technologies give rise to? On 1 March AI Lund invites to a workshop that brings together researchers working with these important questions. The workshop will address the dual questions of why we need human-friendly AI as well as the issues and challenges related to the development and dissemination of these technologies.

Keynote: Anja Bechmann

Professor at Media Studies and director of the interdisciplinary research center DATALAB - Center for Digital Social Research at Aarhus University in Denmark

Bio: Anja Bechmann (she/her) is professor at Media Studies and director of the interdisciplinary research center DATALAB - Center for Digital Social Research at Aarhus University in Denmark. Her research examines digital and social media communication and collective behavior using large-scale data collection and applied machine learning. She is Chief Investigator (CI) of the EU project NORDIS, executive board member of the European Digital Media Observatory (EDMO), and appointed member of the Academy of Technical Sciences. She has been a member of the EU Commission High-level expert group on disinformation and is continously invited as an independent academic to provide research-based talks, white papers or comments on drafts by e.g., European Parliament, Ministry of Defense, and Ministry of Culture. In 2019 she was appointed Thinker in Residence by the Royal Flemish Academy of Belgium for Science and the Arts. Her research has been funded by national and international research councils such as the Danish Council for Independent Research, Swedish Research Council, Danish Agency for Science and Innovation, Horizon 2020, EU CEF, and Aarhus University Research Foundation.

Programme outline

9.30 Registration, fika & mingle

10.15 Session 1: (Why) do we need human-friendly AI?

12.00 Lunch and mingle

13.15 Session 2: Issues and challenges in the development of human-friendly AI

15.00 Fika & mingle

15.30 End of workshop

Speakers

TBA

Registration

To participate is free of charge, but you need to sign up at: https://www.ai.lu.se/2023-03-01/registration

Organisation and contact

• Daniel Møller Ølgaard, Lecturer, Department of Political Science
• Anamaria Dutceac Segesten, Senior Lecturer, European Studies
• Jonas [dot] wisbrant [at] cs [dot] lth [dot] se (Jonas Wisbrant), AI Lund