In a recent Tweet political philosopher Annette Zimmermann presented a meme-like overview of different types of ‘ai ethics papers’. One of those stereotypical papers was the ‘Why ethics (and philosophy more broadly) can’t possible address questions of power’-paper. Indeed, various papers have been published the last couple of years lamenting the lack of ‘power’ (and/or ‘politics’) present in (ethical) analysis of AI. Think, for instance, about an often-quoted paper by Ben Green wherein he argues that data science should be thought of as a form of ‘political action’, forms of data politics as argued for by Ruppert et al, or the ‘agonistic’ political qualities of algorithms. Because politics is presumable something important and worth taking into consideration when studying technology, I thought it would be interesting to have a session on politics and power. I do not expect you to have read the mentioned papers (I really don’t), but to merely think of and reflect on the following (partly overlapping) questions:

• What do politics and power mean for you, your research, and discipline?
• How do you include political or power-dimensions in your work (or not)? Do you think this is important/valuable in the first place?
• How and in what way does your work or your discipline depoliticize issues, or attempt to get rid off power? What would that even mean and is it a problem or not?
• Could it be the case that there’s too much attention to power/politics nowadays in tech development (eg ‘If everything is political, politics becomes meaningless’)?
After discussing these questions together, and if there’s time left, Gijs will discuss several notions of politics and power found in the philosophical literature (probably some Mouffe, Arendt, Foucault, Lukes), which will allow us to compare our discussion with theirs.

The speaker suggests taking a look at the blog entry “Don’t Call AI “Magic” by M.C. Elish on the Data & Society blog, drawn from her collaboration with Danah Boyd, and for those who are keen, a little bit of media history (light and fun reading for skimming) from Sharkey, Noel, and Amanda Sharkey. “Artificial Intelligence and Natural Magic.” Artificial Intelligence Review 25, no. 1 (April 1, 2006): 9–19. https://doi.org/10.1007/s10462-007-9048-z

This talk is in the continuation of the session held on 23 Sep 21.

Prof. George Fletcher will host the session and has proposed the attached article as a starting point for discussion.

Here are two other suggested commentaries:

• Video
• Article

We will be discussing the notion of trust, and how it relates to our respective disciplines’ view of technology. To start thinking about this, here is a suggested reading: a recent essay by Os Keyes on how trust is being used in relation to AI, and how we might want to think more critically about how to do so. Please do invite any colleagues you think might be interested.

The discussion is mainly driven from the chapter “The Concept of Information” authored by Rafael Capurro and Birger Hjørland. The chapter talks about the definition of Information in different areas, their origin and history of the word. Information is a key concept in Sociology, Political Science, Economics, Psychology, Information Science, and several other research areas. The authors raise many interesting questions on Information and discuss them with examples. We will discuss about Information from various perspectives of Data Systems, Social Science and Software Engineering among other questions.

Our topic of discussion will be ‘what is information?’. We will discuss the “The concept of Information” chapter (for which thank you Nadya!) which serves as a starting point for thinking about this problem.