CEPE/IACAP Joint Conference 2021 - Abstract Submission and Registration

Session

Transparency and Explainable AI

29

Not scheduled

80. Beyond informativeness in the epistemology and ethics of XAI in healthcare

Giorgia Pozzi (TU Delft)

Individual Paper

In this presentation, we aim at building upon the framework of what we define the informativeness account of the relation between epistemology and ethics of machine learning (ML). In particular, we refer to the methodology adopted by Mittelstadt et al. (2016). Our first goal is to show that, according to these authors, epistemological issues are instrumental to and independent from...

53. Coping with Black Boxes. The Quest for Epistemic Transparency in Artificial Intelligence

Hajo Greif (Warsaw University of Technology)

Individual Paper

On the background of two contrasting case studies from contemporary Artificial Intelligence, this paper mounts a qualified defence of epistemic opacity in computer modelling and simulation. If epistemic transparency is understood as analytic tractability or the ‘ability to decompose the process between model inputs and outputs into modular steps’ (Humphreys 2004), situations of analytical...