Human Cognitive Bias Mitigation Approaches to Fairness within the Machine Learning Value Chain - A Review and Research Agenda

Aug 6, 2026

Upload by Violet

This systematic review examines how human cognitive biases shape machine learning (ML) systems across the nine phases of the ML algorithmic value chain. Following PRISMA guidelines, it synthesizes 19 studies on the integration and management of bias within ML, focusing on methods that reduce bias and improve fairness. The review highlights gaps including unclear mapping from human cognitive biases to ML biases, missing bias measurement metrics, bias reintroduction during debiasing, and the need for human intervention, informing both theoretical and practical research directions.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages11

Reading Time28 min

AccessFree to download

Abstract

1. Introduction

Research Questions

Dakota State University Beadle Scholar | Research & Publications | College of Business and Information Systems | | --- | --- | | 2025
Human Cognitive Bias Mitigation Approaches to Fairness within the Machine Learning Value Chain: A Review and Research Agenda
Stephen Surles
Dakota State University
Cherie Noteboom
Dakota State University
Omar F. El-Gayar
Dakota State University
Follow this and additional works at: [https://scholar.dsu.edu/bispapers](https://scholar.dsu.edu/bispapers) | | Recommended Citation Surles, Stephen; Noteboom, Cherie; and El-Gayar, Omar F., "Human Cognitive Bias Mitigation Approaches to Fairness within the Machine Learning Value Chain: A Review and Research Agenda" (2025) . Research & Publications. 439. [https://scholar.dsu.edu/bispapers/439](https://scholar.dsu.edu/bispapers/439) This Conference Proceeding is brought to you for free and open access by the College of Business and Information Systems at Beadle Scholar. It has been accepted for inclusion in Research & Publications by an authorized administrator of Beadle Scholar. For more information, please contact [repository@dsu.edu](repository@dsu.edu). Proceedings of the 58th Hawaii International Conference on System Sciences | 2025 Human Cognitive Bias Mitigation Approaches to Fairness within the Machine Learning Value Chain: A Review and Research Agenda Stephen Surles Dakota State University [stephen.surles@trojans.dsu.edu](stephen.surles@trojans.dsu.edu) Cherie Noteboom Dakota State University [cherie.noteboom@dsu.edu](cherie.noteboom@dsu.edu) Omar El-Gayar Dakota State University [omar.el-gayar@dsu.edu](omar.el-gayar@dsu.edu) Abstract This systematic review examines the influence of human cognitive biases on machine learning (ML) systems across the 9 phases of the ML algorithmic value chain. Following the PRISMA guidelines, it synthesizes 19 studies on bias integration and management within ML, highlighting techniques to reduce bias and increase fairness. The review identifies key gaps: the unclear translation of human cognitive biases to ML biases, absence of metrics to measure biases, re-introduction of biases during debiasing, and the critical need for human intervention. These findings prompt several research themes spanning human cognition and algorithmic bias. The theoretical implications are three-fold: extending bias concepts to human cognition, creating an agenda to associate cognitive biases with ML outcomes, and assessing the need for a new or extended discipline. Practically, it raises awareness of human cognition in ML fairness, leading to improved methods for data handling. Keywords: cognitive bias, fairness, discrimination, machine learning, systematic literature review 1. Introduction As humans, we harbor inherent values and biases that are imprinted upon us through all our life learned experiences and interactions (McLarney et al., 2021) . Unfortunately, we are often unaware of these societal biases and how they, in turn, may frame our future cognitive biases. Cognitive biases are described as inherent tendnecies that cause individuals to deviate from logcal or rational decision-making. These biases arise from the brain's reliance on heuristics—mental shortcuts used to simplify complex information processing, which can lead to systematic errors. While these biases may help in making quick decisions in certain situations, they often result in flawed reasoning, especially in uncertain or complex environments (Korteling & Toet, 2022) . Because of this, our very behaviors produce artifacts that exhibit the biases we have (Chouldechova & Roth, 2020) . For example, when negoatiating a salary or the price of a car, the first offer acts as an anchor, strongly influencing the follow-up discussions – a phenomena known as anchoring bias. Or, a person that strongly believes in something may seek out information that confirms the bias they already have – known as confirmation bias. W

FAQs

What does the systematic review analyze about bias in machine learning?

It examines how human cognitive biases influence machine learning systems across the nine phases of the ML algorithmic value chain, and synthesizes studies on integrating and managing bias to improve fairness.

Which methodology and evidence base does the review use?

The review follows PRISMA guidelines and synthesizes 19 studies, summarizing techniques used to reduce bias and increase fairness.

What key research gaps does the review identify?

It identifies unclear translation from human cognitive biases to ML biases, lack of metrics to measure biases, re-introduction of biases during debiasing, and a critical need for human intervention.

Aug 6, 2026

Upload by Violet