Abstract

Generative AI models, such as GPT-4.o, have become increasingly prevalent in various domains, from creative content generation to decision-making support. However, these models are susceptible to biases present in their training data, which can lead to biased or unfair outcomes. While various methods exist for detecting and mitigating bias in AI, there is a lack of standardized, accessible tools for systematically evaluating AI-generated content. Biases in Generative AI input (both implicit and explicit) will result in biased Generative AI output, so it is also important to analyze input. This poster presents a decision tree as a systematic tool to help users identify and evaluate potential biases in generative AI input and output.

Notes

References:

Babaei, G., Banks, D., Bosone, C., Giudici, P., & Shan, Y. (2024). Is ChatGPT More Biased Than You? Harvard Data Science Review, 6(3). https://doi.org/10.1162/99608f92.2781452d

FitzGerald, C., & Hurst, S. (2017a). Implicit bias in healthcare professionals: A systematic review. BMC Medical Ethics, 18(1), 19. https://doi.org/10.1186/s12910-017-0179-8

FitzGerald, C., & Hurst, S. (2017b). Implicit bias in healthcare professionals: A systematic review. BMC Medical Ethics, 18(1), 19. https://doi.org/10.1186/s12910-017-0179-8

Gross, N. (2023). What ChatGPT Tells Us about Gender: A Cautionary Tale about Performativity and Gender Biases in AI. Social Sciences, 12(8), 435. https://doi.org/10.3390/socsci12080435

Description

Generative AI models, such as GPT-4, are prone to perpetuating and amplifying biases present in their training data, raising ethical concerns about fairness and equity in AI-generated outputs. This poster aims to present a comprehensive decision tree that serves as a systematic tool for evaluating generative AI input and output for bias. The decision tree will guide users through the identification and assessment of various types of biases, including gender, racial, and cultural biases.

Author Details

Debra Sullivan, PhD, MSN, CNE, COI, Walden University; Christine Frazer, PhD, CNS, CNE, Walden University

Sigma Membership

Phi Nu

Type

Poster

Format Type

Text-based Document

Study Design/Type

Other

Research Approach

Other

Keywords:

Health Equity, Social Determinants of Health, Virtual Learning, Faculty Development, Emerging Technologies

Conference Name

48th Biennial Convention

Conference Host

Sigma Theta Tau International

Conference Location

Indianapolis, Indiana, USA

Conference Year

2025

Rights Holder

All rights reserved by the author(s) and/or publisher(s) listed in this item record unless relinquished in whole or part by a rights notation or a Creative Commons License present in this item record. All permission requests should be directed accordingly and not to the Sigma Repository. All submitting authors or publishers have affirmed that when using material in their work where they do not own copyright, they have obtained permission of the copyright holder prior to submission and the rights holder has been acknowledged as necessary.

Review Type

Abstract Review Only: Reviewed by Event Host

Acquisition

Proxy-submission

Date of Issue

2025-11-18

Click on the above link to access the poster.

Share

COinS
 

A Decision Tree for Evaluating Generative AI Output and Input for Bias

Indianapolis, Indiana, USA

Generative AI models, such as GPT-4.o, have become increasingly prevalent in various domains, from creative content generation to decision-making support. However, these models are susceptible to biases present in their training data, which can lead to biased or unfair outcomes. While various methods exist for detecting and mitigating bias in AI, there is a lack of standardized, accessible tools for systematically evaluating AI-generated content. Biases in Generative AI input (both implicit and explicit) will result in biased Generative AI output, so it is also important to analyze input. This poster presents a decision tree as a systematic tool to help users identify and evaluate potential biases in generative AI input and output.