Solving AI Epidemiological Data Bias in Indian Country Healthcare

Healthcare technology should improve outcomes for every community, not just the majority.

Artificial intelligence promises personalized medicine, but it often functions as a mirror reflecting deep-seated human prejudices. When algorithms train on historical records lacking sufficient data from Indian Country, they inevitably inherit and automate past disparities. A significant issue identified by researchers is that most U.S. patient data originates from only three states—California, Massachusetts, and New York—completely ignoring the realities of rural Native populations. Consequently, AI models prioritizing cost over care needs have historically favored healthier non-Native patients over Indigenous-identified patients, leading to dangerous misdiagnoses for those in our community.

Indigenous People are a Small Data Set Requiring Surgical Care

We must acknowledge that Indigenous people represent a “small data set,” a reality that demands surgical handling with extreme care rather than broad algorithmic generalizations. Research co-authored by Fay Cobb Payton of Rutgers University, alongside Thelma C. Hurd of Meharry Medical College and Darryl B. Hood of Ohio State University, highlights how algorithms rely on “big data” while missing the critical “small data” of social determinants like transportation or food access. This oversight is devastating for patients relying on the Indian Health Service (IHS). Ellsworth believes securing a seat at the table means pushing for global standards that mandate data reflecting these socioeconomic realities to prevent algorithmic failures.

Engaging Native-serving Practitioners Ensures Culturally Competent AI

The path forward necessitates human intervention to ensure technology serves the intricate web of Native health needs. Healthcare for Native Americans is a multi-layered system that includes federally-run IHS facilities, tribally-managed programs, and Urban Indian health centers, all complemented by Tribal Epidemiology Centers and private hospitals.

To ensure AI contributes meaningfully to our Tribal Citizens, it must be integrated into a broader global healthcare agenda aligned with the principles of respect. This calls for coordinated investment in inclusive data governance, participatory research, and capacity‐building at the urban and rural level. Continuous evaluation, algorithmic accountability, and the involvement of multidisciplinary stakeholders—including Native patients, medical schools, frontline health workers, and policy leaders—are essential to ensuring both technical soundness and ethical legitimacy.

We must integrate insights from practitioners and Native patients to oversee these systems, ensuring they capture the lived experiences of patients navigating these diverse care settings.

As a Native-owned small business (ISBEE), Ellsworth is committed to partnering with tribes to find solutions that mitigate epidemiological data bias. Let us ensure technology advances healthcare fairly for all nations.