[Paper Review] Fairness And Bias in Artificial Intelligence: A Brief Survey of Sources, Impacts, And Mitigation Strategies

Emilio Ferrara|arXiv (Cornell University)|Apr 16, 2023

Ethics and Social Impacts of AISocial Sciences51 references74 citations

TL;DR

This paper surveys the sources, societal impacts, and mitigation strategies for fairness and bias in AI, with emphasis on generative AI bias and the need for interdisciplinary approaches.

ABSTRACT

The significant advancements in applying Artificial Intelligence (AI) to healthcare decision-making, medical diagnosis, and other domains have simultaneously raised concerns about the fairness and bias of AI systems. This is particularly critical in areas like healthcare, employment, criminal justice, credit scoring, and increasingly, in generative AI models (GenAI) that produce synthetic media. Such systems can lead to unfair outcomes and perpetuate existing inequalities, including generative biases that affect the representation of individuals in synthetic data. This survey paper offers a succinct, comprehensive overview of fairness and bias in AI, addressing their sources, impacts, and mitigation strategies. We review sources of bias, such as data, algorithm, and human decision biases - highlighting the emergent issue of generative AI bias where models may reproduce and amplify societal stereotypes. We assess the societal impact of biased AI systems, focusing on the perpetuation of inequalities and the reinforcement of harmful stereotypes, especially as generative AI becomes more prevalent in creating content that influences public perception. We explore various proposed mitigation strategies, discussing the ethical considerations of their implementation and emphasizing the need for interdisciplinary collaboration to ensure effectiveness. Through a systematic literature review spanning multiple academic disciplines, we present definitions of AI bias and its different types, including a detailed look at generative AI bias. We discuss the negative impacts of AI bias on individuals and society and provide an overview of current approaches to mitigate AI bias, including data pre-processing, model selection, and post-processing. We emphasize the unique challenges presented by generative AI models and the importance of strategies specifically tailored to address these.

Motivation & Objective

• Motivate the study of fairness concerns across AI applications (healthcare, employment, criminal justice, credit, GenAI).
• Characterize sources of bias (data, algorithm, human decisions) and define AI bias and generative AI bias.
• Explore societal impacts of biased AI and stereotypes reinforcement.
• Review mitigation strategies (data preprocessing, model selection, post-processing) and ethical considerations.

Proposed method

• Conduct a systematic literature review across multiple academic disciplines.
• Provide definitions and typologies of AI bias, including generative AI bias.
• Discuss mitigation strategies and their ethical implications.
• Highlight challenges and the need for interdisciplinary collaboration to improve effectiveness.

Experimental results

Research questions

• RQ1 What are the primary sources of bias in AI systems (data, algorithms, human decisions)?
• RQ2 What are the societal impacts of biased AI, particularly with generative AI?
• RQ3 What mitigation strategies exist for AI bias, and what ethical considerations do they entail?
• RQ4 How should interdisciplinary collaboration be organized to address fairness and bias in AI?

Key findings

• Biased AI stems from data, algorithmic, and human decision processes.
• Generative AI bias is an emergent issue that can reproduce and amplify societal stereotypes.
• Biased AI can perpetuate inequalities and reinforce harmful stereotypes in various domains.
• Mitigation approaches include data pre-processing, model selection, and post-processing, each with ethical considerations.
• Generative AI presents unique challenges requiring tailored strategies.