AI, Equity, and Accountability: What Every Healthcare Leader Needs to Know Before Deploying AI
Guest Spotlight:
January Montaño, MPS is the CEO and Founder at January's Advisory Group, a Denver-based AI governance consultancy focused on helping healthcare organizations build AI systems that are equitable, auditable, and financially sustainable. With nearly 30 years of experience across healthcare policy, Medicaid, public health, and healthcare operations, she has advised health plans, hospitals, and government agencies on improving outcomes for diverse patient populations.
Her career includes writing Medicaid policy, leading international PPE procurement efforts for Colorado during the COVID-19 pandemic, serving on Governor Jared Polis’ healthcare affordability roadmap, and testifying before the Colorado General Assembly on algorithmic bias in artificial intelligence. January is also the creator of the ROI Equity Framework, a methodology designed to help organizations evaluate both the financial and equity impacts of AI deployment.
Here are the 3 Key Takeaways from our Conversation:
1. AI Is Only as Fair as the Data It Learns From
One of the biggest misconceptions about AI is that it operates objectively. In reality, AI learns from historical healthcare data and that data often reflects longstanding disparities in access, utilization, and outcomes.
As January explained, algorithms can unintentionally use factors such as ZIP codes, healthcare utilization patterns, and socioeconomic indicators as proxies for race and income. The result is that certain populations may be assigned lower risk scores, experience longer wait times, or receive less accurate recommendations not because of their health status, but because of historical inequities embedded in the data.
The lesson: AI does not create bias, but it can rapidly scale existing bias if organizations fail to audit for it.
2. AI Has Tremendous Potential But Governance Determines the Outcome
From identifying rare diseases earlier to reducing physician burnout and improving fraud detection, AI is already delivering meaningful benefits across healthcare.
However, the difference between a successful AI deployment and a harmful one comes down to governance.
January emphasized that healthcare organizations must move beyond asking vendors about efficiency and accuracy. Instead, leaders should demand evidence of equity testing, differential error rate analyses, model validation across diverse populations, and clearly defined accountability structures.
Without these safeguards, organizations may unknowingly deploy tools that create financial, operational, and legal risks while impacting patient care.
3. Accountability for AI Decisions Still Rests with Healthcare Organizations
One of the most important themes from the discussion was liability.
As courts begin examining AI-related healthcare decisions, early cases suggest that health plans and provider organizations, not technology vendors, are being held accountable when AI-driven decisions result in denied care or patient harm.
This creates an urgent need for CEOs, CIOs, compliance leaders, and actuaries to establish robust governance frameworks before implementation. Organizations must know who owns each AI model, how performance is monitored, and what processes exist to identify and correct unintended consequences.
The message is clear: adopting AI without governance is no longer a technology risk it is an enterprise risk.
As AI adoption accelerates across healthcare, the organizations that succeed will not simply be the ones that implement AI the fastest. They will be the ones that deploy it responsibly, transparently, and equitably.
Listen to the full episode of When Health Freezes Over now!
👉 YouTube - https://www.youtube.com/channel/UC7CR1wzokjtVdyyWHLziUJQ/
👉 Spotify - https://open.spotify.com/show/3bRobaBZlM3IbCJ5334PJV?si=41aa6416371e4bc2
👉 Apple Podcast - https://podcasts.apple.com/us/podcast/when-health-freezes-over/id1887501951
#WhenHealthFreezesOver
