Become a Patreon!
Abstract
Excerpted From: Maya C. Jackson, Artificial Intelligence & Algorithmic Bias: the Issues with Technology Reflecting History & Humans, 16 Journal of Business & Technology Law 299 (2021) (168 Footnotes) (Full Document)
In 2015, Google's photo application algorithm was proven flawed after it mistakenly tagged a photo of two Black people as gorillas. The system's algorithm lacked sufficient training with images of darker skin tones. Similarly, in 2017, an algorithm used to create a no-touch soap dispenser was poorly trained to recognized different shades of skin color. As a result, the dispenser only responded to white hands while failing to respond to black and brown hands. These stories, like many others, illustrate the workings of algorithmic bias, a term used to describe systematic and repeatable errors in a computer system that creates unfair and discriminatory practices against various legally protected characteristics like race and gender.
This paper explores how artificial intelligence technologies, such as machine learning and deep learning algorithms, are constructed in ways that create bias and discriminatory outcomes against individuals in various environments, including workplaces and healthcare systems. Specifically, this paper will explore algorithmic bias, analyze how it violates individuals' rights under the Civil Rights Act of 1964 and suggest potential remedies. Section II will provide a descriptive background of algorithms. Section III will then explain algorithmic bias. Section IV will discuss the history of racial and gender discrimination and indicate how it led to algorithm bias today. Section V will describe how algorithmic bias, via human bias or overrepresented or underrepresented data collection, effects today's society in the employment and healthcare realm. Section VI will explore the legal implications on algorithmic bias under the Civil Rights Act of 1964. Finally, Section VII will recommend other remedies to eliminate algorithmic bias.
[. . .]
Technological advances are always known for their contributions to social growth, but it is just as important to be mindful that they are not always perfect or functioning as desired. Although artificial intelligence systems have expanded and improved over the years, the technology may only be as well as the people or data that create it. Thus, it is imperative to properly collect, train and oversee the data periodically to prevent algorithmic bias.
The author is a J.D. Candidate, 2021, at the University of Maryland Francis King Carey School of Law.
Become a Patreon!
