Observability, algorithmic systems and the public domain

Highly complex algorithmic systems increasingly interweave with daily practices in public domains such as education, media, and public management. While these complexities are largely invisible, they replicate biases and assumptions affecting everyday social practices in the environments in which they are operating, exercising unaccountable and potentially harmful influence.

In this one-day symposium, organised by the 'Governing the Digital Society' program from Utrecht University, co-directed by AlgoSoc members José van Dijck, Janneke Gerards, Albert Meijer and Mirko Schäfer, a group of scholars and professionals is brought together to investigate, debate and address the following question:

What conditions, means, and processes of knowledge production can strengthen observability in/of complex algorithmic systems in public domains?

The focus on exploring this question will be on the use and usefulness of the concept of ‘observability’— "a [...] pragmatic way of thinking about the means and strategies to hold platforms accountable" (Rieder and Hofmann, 2020) — for the redesign and redevelopment of complex algorithmic systems and applications. Through the lens of observability, and in reference to concrete cases, the symposium explores "conditions, means, and processes of knowledge production about large-scale socio-technical systems"

10.15 - Walk-in and coffee

10.30 - Keynote "Platform Observability" by Bernhard Rieder. Moderator: José van Dijck

11.30 - Break

11.45 - Algorithmic Controls & Explainability by Dennis Nguyen & Erik Hekman

12.45 - Lunch

13.15 - How Algorithmic Accountability Succeeds or Falls Short in Practice by Maranke Wieringa

14.15 - Break

14.30 - API Governance: The Case of Facebook’s Evolution by Anne Helmond & Fernando van der Vlist

15.15 - Practices of Observing in Different Domains. Moderators: Karin van Es & Niels Kerssens

16.00 - Drinks