Addressing Discrimination in Artificial Intelligence
Mitigate the risk of automating human bias
May 12, 2020
15 Minute Read
Artificial intelligence (AI) is already transforming the health care industry. But as with any new technology in health care, there is a great deal of fear and mistrust around the use of AI. One ethical risk gets significant attention: the potential for algorithms to learn, automate, and scale institutionalized biases.
There is legitimate cause for concern. Bias already exists in many internally built or commercially sold algorithms used to make care decisions for millions of patients. Overlooking the existence of bias leads to irresponsible applications of AI. But avoiding AI altogether prevents us from realizing the benefits of this technology. When applied rigorously and with health equity in mind, AI can help us find and fix flaws in the current system. Health care leaders must deliberately develop AI solutions centered around health equity.