Acknowledgments and Final Postings

I would first like to thank my expert advisers Mr. Govindarajulu and Mr. Zambello for taking the time to be my expert advisers this year. My research project would not have been as successful as it was without their support and feedback. I would like to express my appreciation to Mr. Govindarajulu for his suggestion on websites where I can find the datasets for my research and the data analysis tools to analyze the datasets. I would also like to thank Mr. Zambello for suggesting areas to research in, such as the use of AI in the COVID-19 pandemic.

Next, I would like to thank my parents for their continued support throughout this research process. I really appreciated their interest in my research project and encouragement throughout my study.

I would also like to thank my classmates who took AP Research with me. Their thoughtful feedback on my blog posts, paper, and presentations helped me to improve my work.

Lastly, I would like to express my very great appreciation to my supervisor, Mr. Winkelman. His constant encouragement and support, as well as his perseverance through a new class, helped keep me motivated to push through many of the challenges I faced. I am extremely thankful for his commitment to our class's success.

Many researchers believe that artificial intelligence (AI) could be a solution to addressing the lack of proper healthcare in developing countries. Studies have found that AI can help with diagnosing a patient and keeping track of medical records. Nevertheless, AI isn’t perfect as research has found there to be racial bias in some AI algorithms in the United States, caused by unrepresentative data. These racially biased algorithms can lead to harmful effects, especially for minorities. This study aims to investigate if medical datasets, that can be used in AI algorithms, are diverse enough to be effective in healthcare in developing countries.

To determine if medical datasets have sufficient racial diversity to be effective in developing countries, I analyzed four medical datasets to conclude if each racial group was sufficiently represented. Using a chi-square goodness of fit test I determined if the sample data matched the population. The results showed that three out of the four datasets analyzed did not represent minorities sufficiently and did not match the population. The results suggest that the majority of healthcare datasets do not represent minorities adequately which indicates that algorithms that use these datasets, will not be as effective in healthcare in developing countries.