United Regional Health Care System

Ali Booher is a healthcare executive at United Regional Health Care System, passionate about operational excellence, quality, and digital transformation in the form of advancing equity through responsible innovation.

Artificial Intelligence (AI) is transforming healthcare—from streamlining diagnostics to enhancing patient engagement and optimizing operations. However, as AI accelerates across the sector, it brings complex ethical, legal and policy challenges that demand thoughtful, inclusive and equity-driven solutions.

The Dual Challenge: Innovation and Oversight

The promise of AI in healthcare is well-documented. It improves diagnostic accuracy, personalizes treatment and alleviates clinician burnout. Yet, rapid integration also raises concerns around bias, liability and access. Without adequate oversight, AI could entrench disparities, erode trust and harm patient safety. A hybrid governance approach is needed—one that encourages responsible innovation while protecting vulnerable populations.

AI in healthcare raises several pressing concerns. In clinical decision-making and liability, there is an ongoing debate over whether AI is advisory or authoritative and critically who is liable when AI-driven results harm? Algorithm bias is another risk as AI may replicate or amplify systemic inequities if datasets lack diversity. Strong data governance is essential to preserve patient data privacy, security and consent without stifling innovation. Finally, access and equity must be addressed, since social determinants of health (SDOH) such as income, education and digital literacy shape AI's real-world utility and fairness.

Stakeholder Values and Competing Incentives

Every stakeholder—patients, providers, health systems, insurers, regulators and tech developers—views AI through a distinct lens:

• Patients want privacy, accuracy, and transparency.

• Providers need liability protections and clinical clarity.

• Health systems prioritize efficiency and cost control.

• Insurers seek risk prediction and automation.

• Regulators balance innovation with safety.

• Developers favor speed, adoption, and IP protection.

Aligning these priorities requires clear policy frameworks grounded in transparency, accountability and equity.

Policy Recommendations

A national hybrid governance model can provide the structure needed to guide ethical AI adoption. Recommended pillars include:

1. Mandatory Bias Audits—Regularly test algorithms for racial, gender, and socioeconomic bias.

2. Transparency Standards—Require explainability and auditability in clinical decision-support tools.

3. Equity Incentives—Fund AI deployment in underserved communities and support digital literacy.

4. Workforce Development—Train providers to use AI ethically and effectively.

Legislative and Administrative Strategy

The path forward involves both congressional legislation and rulemaking through agencies like the FDA, CMS and AHRQ. The FDA can expand SaMD (Software as a Medical Device) oversight. CMS should modernize reimbursement for AI tools. AHRQ could lead audits and evaluation science.

“AI is only as just and effective as the humans who create and implement it. Equity must be embedded from data collection to deployment.”

Rulemaking under the Administrative Procedure Act (APA) will ensure public engagement and stakeholder input. Continuous policy evaluation using frameworks like RE-AIM will track implementation, adoption and outcomes.

Implementation Science and Real-World Learning

Data-driven iteration is key. Measuring AI’s impact on diagnostic accuracy, bias reduction and access equity will inform refinements. Evaluation science ensures adaptive, context-sensitive policy execution.

The Political and Social Landscape

Current political polarization complicates national consensus. Debates over data privacy, government overreach and equity mandates may stall progress. Meanwhile, SDOH create tangible barriers—digital deserts, housing instability and educational gaps all hinder AI adoption and exacerbate disparities.

Culturally sensitive policymaking, community engagement and targeted funding are vital to overcoming resistance and ensuring inclusive implementation.

The Trade-Offs We Must Manage

• Transparency vs. IP Protection—Balancing explainability with proprietary development.

• Speed vs. Safety—Ensuring rigorous validation without stifling innovation.

• Cost Savings vs. Workforce Stability—Preventing job displacement and burnout.

Conclusion: Designing AI for Justice and Impact

AI is only as just and effective as the humans who create and implement it. Equity must be embedded from data collection to deployment. As we shape policy, we must reject one-size-fits-all solutions and instead pursue inclusive frameworks that reflect the complexities of our healthcare system.

By embracing thoughtful governance and empowering every stakeholder—especially those historically marginalized—we can ensure AI becomes a catalyst not only for efficiency, but for justice. In the end, our goal should be a future where technology uplifts every patient, protects every provider and transforms healthcare for good.