Opacity in AI: Analysis, Impact, Policy (EASI & IC)

Project: First tier

Project Details


EAISI & Irène Curie project, “Opacity in AI: Analysis, Impact, Policy”
In recent years, philosophers have shown an increasing interest in AI. One set of philosophical questions is due to the fact that AI systems often produce an output where it is not clear why this output was produced; the system is ‘opaque’ (non-transparent) to the subjects, users, and even to the makers of the system. This raises a large number of pressing research questions, including: 1) What, exactly, is opacity in AI? 2) What is the impact of opacity and what is its ethical relevance? 3) Should the impact of opacity be reduced? 4) Can the impact of opacity be reduced, e.g. through “explainable AI”? 5) Can we have data protection or unbiased systems if they are opaque? We welcome research projects on these or related issues.
Short titleOpacity in AI
Effective start/end date1/01/2131/12/24