A Question Bank to Assess Inclusive AI

Abstract

This document is a supplementary resource to the article "A question bank to assess inclusive AI" (DOI: 10.1007/s43681-026-01068-5), accepted in journal of AI and Ethics (Springer). This resource was developed by the CSIRO Data61 AI Diversity and Inclusion Team. The accompanying peer-reviewed paper was led by Dr Rifat Ara Shams (Monash University, formerly affiliated with CSIRO's Data61) with Professor Didar Zowghi and Dr Muneera Bano (CSIRO's Data61) as co-authors and corresponding contacts for the Question Bank resource. It presents a sample of 25 questions drawn from the full Question Bank to Assess AI Inclusivity, a comprehensive framework of 253 questions developed by CSIRO's Data61 AI Diversity and Inclusion Team. The full Question Bank is organised across five pillars: Human (57 questions) — stakeholder engagement, diversity practices, AI literacy, and community inclusion Data (53 questions) — data governance, privacy, Indigenous data sovereignty, and bias in labelling Process (92 questions) — pre-development, development, and post-deployment inclusive AI practices System (23 questions) — holistic evaluation of AI systems including values, design, and institutional alignment Governance (28 questions) — biometric policies, risk management, regulatory compliance, and responsible AI leadership The Question Bank is designed to help AI practitioners, researchers, and organisations systematically assess and improve diversity, equity, and inclusion (DEI) practices throughout the AI lifecycle, from pre-design through to post-deployment monitoring. For access to the full Question Bank or reuse permissions, please contact the corresponding authors: Didar Zowghi (didar.zowghi@csiro.au) or Muneera Bano (muneera.bano@csiro.au), CSIRO's Data61.

Ähnliche Arbeiten