Protocols not Platforms: The Case for Human-Centered Explainable AI

Authors

  • Michael Ridley University of Guelph

DOI:

https://doi.org/10.29173/cais1859

Abstract

Currently explainable AI (XAI) is embedded in large, centralized platforms such as Facebook, Google or TikTok. These platforms control the nature and extent of explanations for their recommendations, decisions, and predictions raising the possibility of manipulation or deception. Human centered XAI (HCXAI) promotes explanatory systems not merely explanations as part of a set of principles supporting the non-expert, lay public. This position paper proposes moving from platform enabled HCAXI to protocol based HCXAI to facilitate user focused, independent explanatory systems more conducive to building and sustaining user trust and system accountability.

Downloads

Published

2024-09-10

How to Cite

Ridley, M. (2024). Protocols not Platforms: The Case for Human-Centered Explainable AI. Proceedings of the Annual Conference of CAIS Actes Du congrès Annuel De l’ACSI. https://doi.org/10.29173/cais1859

Issue

Section

Articles