by Sheila Thorne
Artificial Intelligence (AI) emerged as a formal field in 1956 at the Dartmouth Summer Research Project at Dartmouth College in New Hampshire. After decades of advances and setbacks, AI is now valued for processing large datasets, learning patterns quickly, and supporting decision-making.
- Advertisement -
AI adoption in U.S. healthcare is accelerating at warp speed to address inefficiencies, workforce shortages, and uneven outcomes. In 2023, healthcare AI investment reached $1.4B, with organizations using AI to automate administrative work, streamline clinical operations, and improve outcomes—especially for underrepresented and underserved patients.
AI is reshaping pharmaceutical clinical research by accelerating drug discovery and development. Where one new drug often requires 10–15 years and about $2.2B with no guarantee of approval, AI methods may reduce
timelines and costs substantially—helping deliver therapies sooner, including for conditions that disproportionately affect marginalized communities of color such as heart disease, cancer and infectious diseases.
However, major risks remain: health outcomes for people of color are poor—even among insured and middle-class patients. A key contributor is limited representation in AI development: fewer than 20% of practitioners are women and fewer than 2% are people of color. Clinical evidence has also been historically skewed toward white, urban, middle-aged men, in North America and then generalized to other groups, leaving dangerous gaps in diagnosis and treatment. Some AI algorithms do not recognize melanated skin.
AI models trained on incomplete, biased, or unrepresentative data can reproduce and amplify disparities. The tech adage “Garbage in, garbage out” (GIGO) applies: weak inputs yield unreliable outputs, which can lead to misleading treatment recommendations and further embed inequities in care.
To realize healthcare AI’s benefits across populations, priorities include:
- Diversify AI teams: Recruit and advance women and people of color in AI research, development, and leadership.
- Build representative datasets: Ensure clinical trials reflect U.S. diversity, with targeted recruitment of underrepresented groups.
- Engage communities: Partner with community organizations to build trust and align solutions to real needs.
- Improve transparency: Document data sources, limitations, and bias risks; set clear accountability for outcomes.
- Share what works: Publish case studies where AI reduces health disparities or improves health outcomes for marginalized groups.
AI is revolutionizing healthcare, but benefits will be limited without inclusive design, equitable data, and accountability. Prioritizing representative evidence and transparent practices can help healthcare AI improve outcomes for all—especially populations historically left behind.
Sheila Thorne / President & CEO / Multicultural Healthcare Marketing Group, LLC has attended a dozen AI seminars including being a delegate at 2025 International AI Summit in Cape Town, South Africa sponsored by the University of the Western Cape. Visit sheilathorne.com for more information.
