Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

AI for All: Identifying AI incidents Related to Diversity and Inclusion

Rifat Ara Shams, Didar Zowghi, Muneera Bano

TL;DR

This study addresses the challenge of embedding Diversity and Inclusion (D&I) in AI by systematically analyzing real-world AI incidents from AIID and AIAAIC. It introduces a four-condition decision tree to identify D&I-related incidents and validates it through card sorting and focus groups, culminating in a publicly accessible repository of D&I-incidents. Results show substantive shares of incidents linked to D&I, with race, gender, and age being the most frequent attributes affected. The work provides a methodological blueprint for ongoing, open analysis of AI harms related to D&I, offering practical guidance for researchers, practitioners, and policymakers to promote inclusive and fair AI systems, and outlines a roadmap for future expansion and automation.

Abstract

The rapid expansion of Artificial Intelligence (AI) technologies has introduced both significant advancements and challenges, with diversity and inclusion (D&I) emerging as a critical concern. Addressing D&I in AI is essential to reduce biases and discrimination, enhance fairness, and prevent adverse societal impacts. Despite its importance, D&I considerations are often overlooked, resulting in incidents marked by built-in biases and ethical dilemmas. Analyzing AI incidents through a D&I lens is crucial for identifying causes of biases and developing strategies to mitigate them, ensuring fairer and more equitable AI technologies. However, systematic investigations of D&I-related AI incidents are scarce. This study addresses these challenges by identifying and understanding D&I issues within AI systems through a manual analysis of AI incident databases (AIID and AIAAIC). The research develops a decision tree to investigate D&I issues tied to AI incidents and populate a public repository of D&I-related AI incidents. The decision tree was validated through a card sorting exercise and focus group discussions. The research demonstrates that almost half of the analyzed AI incidents are related to D&I, with a notable predominance of racial, gender, and age discrimination. The decision tree and resulting public repository aim to foster further research and responsible AI practices, promoting the development of inclusive and equitable AI systems.

AI for All: Identifying AI incidents Related to Diversity and Inclusion

TL;DR

This study addresses the challenge of embedding Diversity and Inclusion (D&I) in AI by systematically analyzing real-world AI incidents from AIID and AIAAIC. It introduces a four-condition decision tree to identify D&I-related incidents and validates it through card sorting and focus groups, culminating in a publicly accessible repository of D&I-incidents. Results show substantive shares of incidents linked to D&I, with race, gender, and age being the most frequent attributes affected. The work provides a methodological blueprint for ongoing, open analysis of AI harms related to D&I, offering practical guidance for researchers, practitioners, and policymakers to promote inclusive and fair AI systems, and outlines a roadmap for future expansion and automation.

Abstract

The rapid expansion of Artificial Intelligence (AI) technologies has introduced both significant advancements and challenges, with diversity and inclusion (D&I) emerging as a critical concern. Addressing D&I in AI is essential to reduce biases and discrimination, enhance fairness, and prevent adverse societal impacts. Despite its importance, D&I considerations are often overlooked, resulting in incidents marked by built-in biases and ethical dilemmas. Analyzing AI incidents through a D&I lens is crucial for identifying causes of biases and developing strategies to mitigate them, ensuring fairer and more equitable AI technologies. However, systematic investigations of D&I-related AI incidents are scarce. This study addresses these challenges by identifying and understanding D&I issues within AI systems through a manual analysis of AI incident databases (AIID and AIAAIC). The research develops a decision tree to investigate D&I issues tied to AI incidents and populate a public repository of D&I-related AI incidents. The decision tree was validated through a card sorting exercise and focus group discussions. The research demonstrates that almost half of the analyzed AI incidents are related to D&I, with a notable predominance of racial, gender, and age discrimination. The decision tree and resulting public repository aim to foster further research and responsible AI practices, promoting the development of inclusive and equitable AI systems.
Paper Structure (18 sections, 9 figures, 2 tables)

This paper contains 18 sections, 9 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Background and Related Work
  3. Diversity and Inclusion in AI
  4. AI Incidents Related to Diversity and Inclusion
  5. Methodology
  6. Data Collection
  7. AI Incident Databases
  8. Participants of Card Sorting and Focus Groups
  9. Protocol of Card Sorting
  10. Protocol of Focus Groups
  11. Data Analysis
  12. Results
  13. RQ1: Identifying D&I Issues in AI Incidents
  14. RQ2: Extent to Which the Existing AI Incidents are Related to D&I issues
  15. Discussion and Implications
  16. ...and 3 more sections

Figures (9)

  • Figure 1: An overview of the research method
  • Figure 2: Decision tree to detect D&I-related AI incidents
  • Figure 3: Illustration of the functioning procedure of the decision tree to understand its relevance to diversity and inclusion with some examples.
  • Figure 4: The ratio of diversity attributes in AI incidents from AIID database
  • Figure 5: The ratio of diversity attributes in AI incidents from AIAAIC database
  • ...and 4 more figures