Press release: How political is artificial intelligence?

Contact for content-related questions

PD Dr. Andreas Sudmann

Institute for Media Studies
Ruhr-University Bochum
E-Mail: andreas.sudmann@rub.de
Telefon: +49 151 23 01 17 99

Contact CAIS Funding Programs/Press

Dr. Matthias Begenat

Center for Advanced Internet Studies (CAIS) GmbH
Head of Knowledge Transfer and Public Relations
Universitätsstr. 104
D-44799 Bochum
Tel. +49 234 544 96 047
E-Mail: matthias.begenat@cais.nrw

Press release: How political is artificial intelligence?
CAIS sponsors international conference in Bochum

Bochum, 5 September 2018

The conference “The democratization of artificial intelligence: net politics in the age of learning algorithms” will take place in Bochum from 6-7 September. German researchers will discuss at the conference the political implications of using artificial intelligence with colleagues from Canada, Finland, and England. The conference is sponsored by the Center for Advanced Internet Studies, and organized by Andreas Sudmann.

Artificial intelligence (AI) has long been used to identify potential terrorists, to outline political speeches in a simple form, and to evaluate the behavior of users on social media platforms for marketing purposes. Not least the numerous advances in the field of AI have led to growing discussion not only of the technical and economic aspects involved in using the technology, but also of the political questions that doing so raises.

German researchers will meet colleagues from Canada, Finland, and England at the international conference on “The democratization of artificial intelligence” at the European Education Center in Bochum on 6-7 September.

Keynote speaker at the conference will be Roland Memisevic, chief scientist of the German-Canadian company TwentyBN, which develops artificial intelligence and trains AI using not still images or text data as learning material, but videos. “Computers can learn from videos that the world is three-dimensional, that there are objects in the world, and that these objects are subject to a range of physical laws”, says Memisevic of the approach developed by his company. As a core idea, he describes the more efficient handling of large amounts of data. By means of so-called transfer learning, his system does not have to be retrained for each new task, but should be able to use learning experiences for other problems, too.

Andreas Sudmann, who is organizing the conference, builds a bridge to the political properties of artificial intelligence: “Even if in this case it is not part of TwentyBN’s work area, it is true that other systems are used, for example, in the health sector or for monitoring sensitive spaces. The technology is by no means as neutral as it may at first appear, and the decisions that are made there in a fraction of a second also have relevant social effects and implications”. Researchers from various disciplines will discuss the political significance and implications of the developments and use of AI technologies. The language used at the conference is English.

» Further information on the conference and the program