Author: Helen Santoro
Publisher: Scope
Publication Year: 2022
Summary: The following article discusses how artificial intelligence (AI) has proven potent in analyzing historical medical data, chemical compound data, and using it to predict viable treatments. As a result, researchers are advocating for further implementation of machine learning in drug development, which they believe will increase speed and efficiency of medication research. Key figures in the world of drug development have called for the urgent need of ethical frameworks, as the problems of unethical data practices are real and hurtful to companies and consumers alike. Currently, AI is effective in scouring genomic databases to see what factors associate with disease or examining patterns in patients to anticipate how they will respond to a given treatment. However, this is problematic, as genomic databases and old medical records contain information primarily on people of European descent. This biases model efficacy in favor of European descendants, leaving people of other backgrounds out of the equation. Even medical records suffer this problem, while also posing a risk to patient privacy. To remedy this, Santoro notes that ethics need to be incorporated in data analytics from the very beginning of the process. It is not sufficient to simply “sprinkle” ethical principles on top of a project at the end of its life. To do this, analysts need to consider the worst-case outcomes of their projects — essentially, how their models can be used to hurt others, should they be utilized maliciously.