Main Library
Algorithms play a significant role in our daily lives — they unlock our phones, tell us which movies to watch and dictate the content we see on social media. These calculations are also frequently used by healthcare providers to inform diagnosis and treatment plans, so it’s important to remember that algorithms are only as good as the data used to train them.
Much of the data used to train today’s healthcare algorithms reflects the structural racism embedded in the U.S. healthcare system, creating a bias that negatively affects health outcomes among already marginalized populations. That’s why the engineers building healthcare algorithms should move away from using race as a measure of health disparities and pivot to using social determinants of health, according to the speakers featured on Philadelphia Alliance for Capital and Technologies’ Monday virtual panel on medical AI bias.