The power and pitfalls of AI: creating a just, equitable society with technology

November 07, 2023

Graphic representation of ones and zeroes

By Karen Shih

From developing personalized learning platforms tailored to different abilities to identifying patterns and creating interventions to decrease homelessness, artificial intelligence (AI) can be used to transform lives across the United States and around the world.

“AI has the potential to contribute toward equity and racial justice across social policy domains, such as housing, the environment, and health care,” says Ezra Tefera, MS GHPM’22, program director of the Racial Justice x Tech Policy (RJxTP) program in the Institute for Economic and Racial Equity.

A Medical Doctor With a Global Outlook

A medical doctor in his home country of Ethiopia, Tefera had seen firsthand how unequal access to health care caused people to suffer from preventable diseases like cholera, malaria, and HIV, which led him to enroll at Heller with the goal of making a broader impact on disadvantaged communities.

“I became really intrigued by AI’s potential,” Tefera says. “I saw how analytics could be harnessed to revolutionize and build a more equitable health system. It promised to improve accuracy, automate tasks, and reduce human errors associated with medical malpractice, all thanks to AI’s vast amount of computing power.

“However, I also recognized the need to tread cautiously,” Tefera adds. “AI can inadvertently or consciously perpetuate existing biases and disparities if not implemented responsibly and equitably.”

For example, he understood that electronic health records (EHR) reflect the existing use of health care, mostly by people who have insurance. But marginalized communities that have been excluded due to financial barriers, redlining, and housing discrimination, or other systemic issues, aren’t reflected in that data. If policymakers combine this EHR information with ZIP codes to determine health care needs, it can perpetuate and amplify racial inequities.

Now, through the RJxTP program, which takes a multipronged approach to address algorithmic bias, Tefera is working to bring together diverse researchers to tackle the impact of technology on social and health policy issues. Using one of the largest EHR data sets in the country from Sutter Health, the team’s goals include offering anti-racism recommendations, combating algorithmic oppression, and promoting equitable care.

Tackling Bias

Launched in early 2022, RJxTP aims to build diversity and inclusivity in the STEM workforce through co-creating capacity; increase tech policy and racial justice education through micro-credentials, webinars, workshops, and other services; and transform interdisciplinary research into actionable policy.

Five research projects in the program’s incubator are focusing initially on algorithmic discriminations and biases in education, mortgage lending, health, environment, and the criminal-legal system. Tefera is involved in several of these efforts, including one that examines birth inequities in Black birthing persons, who face mortality rates three times higher than their white counterparts.

In addition, he is part of a team developing a priority-setting framework for health care sites by analyzing existing guidelines, regulations, and medical frameworks. For example, present-day AI tools are widely used in clinical practices for triaging patients, deciding who is admitted and seen first by a physician. One of his first deliverables will be an inequity index tool, which will flag potential biases in tech used by health care providers.

“I hope to contribute to a future where tech works in harmony with health and social policy to improve outcomes for all, irrespective of backgrounds or socioeconomic status ... where equity, inclusion, and technology can go hand in hand,” Tefera says.