Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

15 Years of Algorithmic Fairness -- Scoping Review of Interdisciplinary Developments in the Field

Daphne Lenders, Anne Oloo

TL;DR

This scoping review analyzes 1570 algorithmic fairness papers across 15 years from four databases to map how the field has evolved along domain, demographic, and technological lines, with a cross-disciplinary lens on Computer Science and Law. The study reveals a growing domain-specific focus, increasing attention to intersectional demographics, and a shift beyond classification toward a wider set of AI technologies, while still showing a North America–Europe concentration and rising but insufficient Global South representation. By detailing institutional and geographic contexts, the work highlights opportunities for broader interdisciplinary collaboration and more regionally diverse case studies, informing researchers, policymakers, and industry on where biases arise and how to tailor governance. Overall, the findings argue for a more inclusive, domain-aware, and globally representative research agenda to address algorithmic discrimination in real-world deployments. The work also demonstrates the value of integrating legal and technical perspectives to surface regulatory gaps and practical bias mitigation strategies that reflect diverse social contexts.

Abstract

This paper presents a scoping review of algorithmic fairness research over the past fifteen years, utilising a dataset sourced from Web of Science, HEIN Online, FAccT and AIES proceedings. All articles come from the computer science and legal field and focus on AI algorithms with potential discriminatory effects on population groups. Each article is annotated based on their discussed technology, demographic focus, application domain and geographical context. Our analysis reveals a growing trend towards specificity in addressed domains, approaches, and demographics, though a substantial portion of contributions remains generic. Specialised discussions often concentrate on gender- or race-based discrimination in classification tasks. Regarding the geographical context of research, the focus is overwhelming on North America and Europe (Global North Countries), with limited representation from other regions. This raises concerns about overlooking other types of AI applications, their adverse effects on different types of population groups, and the cultural considerations necessary for addressing these problems. With the help of some highlighted works, we advocate why a wider range of topics must be discussed and why domain-, technological, diverse geographical and demographic-specific approaches are needed. This paper also explores the interdisciplinary nature of algorithmic fairness research in law and computer science to gain insight into how researchers from these fields approach the topic independently or in collaboration. By examining this, we can better understand the unique contributions that both disciplines can bring.

15 Years of Algorithmic Fairness -- Scoping Review of Interdisciplinary Developments in the Field

TL;DR

This scoping review analyzes 1570 algorithmic fairness papers across 15 years from four databases to map how the field has evolved along domain, demographic, and technological lines, with a cross-disciplinary lens on Computer Science and Law. The study reveals a growing domain-specific focus, increasing attention to intersectional demographics, and a shift beyond classification toward a wider set of AI technologies, while still showing a North America–Europe concentration and rising but insufficient Global South representation. By detailing institutional and geographic contexts, the work highlights opportunities for broader interdisciplinary collaboration and more regionally diverse case studies, informing researchers, policymakers, and industry on where biases arise and how to tailor governance. Overall, the findings argue for a more inclusive, domain-aware, and globally representative research agenda to address algorithmic discrimination in real-world deployments. The work also demonstrates the value of integrating legal and technical perspectives to surface regulatory gaps and practical bias mitigation strategies that reflect diverse social contexts.

Abstract

This paper presents a scoping review of algorithmic fairness research over the past fifteen years, utilising a dataset sourced from Web of Science, HEIN Online, FAccT and AIES proceedings. All articles come from the computer science and legal field and focus on AI algorithms with potential discriminatory effects on population groups. Each article is annotated based on their discussed technology, demographic focus, application domain and geographical context. Our analysis reveals a growing trend towards specificity in addressed domains, approaches, and demographics, though a substantial portion of contributions remains generic. Specialised discussions often concentrate on gender- or race-based discrimination in classification tasks. Regarding the geographical context of research, the focus is overwhelming on North America and Europe (Global North Countries), with limited representation from other regions. This raises concerns about overlooking other types of AI applications, their adverse effects on different types of population groups, and the cultural considerations necessary for addressing these problems. With the help of some highlighted works, we advocate why a wider range of topics must be discussed and why domain-, technological, diverse geographical and demographic-specific approaches are needed. This paper also explores the interdisciplinary nature of algorithmic fairness research in law and computer science to gain insight into how researchers from these fields approach the topic independently or in collaboration. By examining this, we can better understand the unique contributions that both disciplines can bring.
Paper Structure (14 sections, 8 figures, 1 table)

This paper contains 14 sections, 8 figures, 1 table.

Table of Contents

  1. Introduction
  2. Related Literature & Motivation
  3. Methodology
  4. Databases & Search Query
  5. Selection of Papers
  6. Data Extraction
  7. Results
  8. RQ1: The Need for Domain-Specific Approaches
  9. RQ2: Making Diverse (Intersectional) Demographic Groups the Focus of the Research
  10. RQ3: Moving beyond Classification
  11. RQ4: Considering Global Perspectives
  12. Authors' Affiliation
  13. Geographical Focus in Papers' Content
  14. Discussion & Conclusion

Figures (8)

  • Figure 1: The Web of Science search query to capture relevant literature, based on key phrases in papers' title and abstract
  • Figure 2: The domain focus of papers over the years
  • Figure 3: The domain focus of papers split by first author's expertise
  • Figure 5: The demographic focus of papers over the years
  • Figure 6: The demographic focus of papers split by first author's expertise
  • ...and 3 more figures