Abstract
As AI increasingly automates life-altering decisions, evaluating bias in its decision-making processes becomes crucial. With the growth of data analytics, the use of AI in a variety of sectors has also expanded. AI analytics is currently being used in the fields of education, healthcare, banking, hiring, and other critical sectors. Thus, the inherent biases that these systems present can be crucial to evaluating the best practices to employ fairness in data analytics. These biases may seem insignificant as it is easy to see data as a purely statistical analysis of the truth of any situation. Yet, this view often neglects the ability of statistics to distort the truth to align with specific objectives. A prime example of this bias is evident in the housing sector. Historically, minority groups in the U.S. housing market have been denied loans at a higher rate than their majority counterparts, even when all other factors remain unchanged. Analytical models may interpret this as evidence that minorities pose a higher risk for loans, based on biased data from previous loan denials. Thus, this paper seeks to answer a fundamental question: How can AI-driven decision-making processes be systematically evaluated and improved to ensure fairness while mitigating bias in real-world applications? By analyzing the leading fairness-based algorithms in both ethical considerations and real-world applications, we can address their current shortcomings. In the context of an AI-driven world, research into ethical AI is essential to ensure fairness. Ultimately, ensuring fairness in AI is not just a technical challenge, but a moral imperative to create more equitable and accountable systems.
Language: English
Published on: Mar 20, 2025
Published by: University of North Georgia
In partnership with: Paradigm Publishing Services
Keywords:
© 2025 Annah Fenchak, published by University of North Georgia
This work is licensed under the Creative Commons Attribution 4.0 License.