Privacy. Surveillance. Influence. Medical expert Matthew DeCamp and Colorado State University faculty explored the constellation of issues surrounding artificial intelligence during the annual Provost’s Ethics Colloquium this week.
DeCamp, a medical doctor and associate professor at the Center for Bioethics and Humanities at the University of Colorado Anschutz Medical Campus, was the keynote speaker at the Provost’s Ethics Colloquium on Nov. 28, which also included a panel discussion featuring CSU faculty members conducting AI research.
DeCamp’s keynote at the Lory Student Center included a historical overview of AI and its impact on health care, examining the ethical issues involving facial recognition, chatbots and large public datasets, among other areas.
DeCamp touched on how dermatology software for identifying skin conditions can have performance issues in helping people with darker skin tones and those who identify as women.
He also shared how an algorithm designed to allocate health resources falsely concluded Black patients needed less health care. He explained that the bias occurred because the algorithm used health costs as a proxy for health needs, noting that less money is spent on Black patients because of historical injustices and structural racism.
“You can sometimes have biased artificial intelligence, even when the particular variable in question by race is taken out,” he said. “So, you might think, I’ll make a fair algorithm. I’ll just take race and other obsessional variables out of the algorithm. Well, guess what? AI is so smart that it can eventually figure out biases and replicate them, and perpetuate them, even if you’ve taken those steps to take out race.”