Algorithmic Bias in Medicine: A Study of Discrimination Against PoC in Healthcare AI (PDF)

In the 19th century, those controlling the steel industry amassed considerable fortunes. Subsequently, the 20th century witnessed oil as the principal resource for accumulating wealth. Presently, data, often referred to as the new oil, are among the most valuable assets. Seven of the world's leading corporations, as measured by market value, were involved in data and artificial intelligence (AI) towards the end of 2019. This growth in Big Data and AI has spurred the fourth industrial revolution.
Within this thesis, AI refers to the capacity of machines to execute tasks autonomously, based on algorithms, whilst adaptively responding to unfamiliar circumstances. As such, the thesis examines methods that empower computers to tackle problems that necessitate intelligence when resolved by humans. AI encompasses the study of intelligent problem-solving behaviour and the development of intelligent computer systems. Additionally, the emphasis here is placed on data-driven AI instead of rule-based AI.
The central issue addressed in this thesis pertains to the discrimination faced by People of Colour (PoC), exacerbated by healthcare AI and algorithms. As detailed in the subsequent chapter, research has demonstrated that PoC are treated differently by medical technology; discrimination based on the algorithmic use of ethnic factors is a pervasive problem in hospitals.