AI in Healthcare Series: AI as Your Personal Health Partner

Key Concepts: AI in Healthcare, Personal Healthcare Management, Productivity Gains, Patient Comfort with AI, Healthcare System Transformation, Physician Shortage, Access Issues, Preventive Care, Individualized LLMs, Physician Augmentation, Value-Based Care, Policy and Regulation, FDA Oversight, Experimental Sandboxes, Administrative Efficiency, Clinical AI, Revenue Cycle Management (RCM), Coding Optimization, Ambient Scribes, Healthcare Costs, AI Arms Race, Deflationary Forces, Empowered Patient, Bottoms-Up Disruption.

The Transformative Impact of AI on Personal and Professional Healthcare

The discussion opens with Dr. Syed Mohiuddin, Chief AI Transformation and Strategy Officer at UnitedHealth Group, sharing his personal experience with AI in managing health queries. He notes a 5-10x increase in the volume of personal health queries he addresses, as AI helps answer "lower hanging" questions that previously lingered. Concurrently, his outreach to specialists has decreased by 50%, leading to an estimated 10-20x personal productivity gain. While AI is highly effective for quick checks and "homework" before consultations, it currently cannot replace human expertise for urgent, rare, or geographically specific medical needs, such as finding a specialist for a rare cancer, which still requires direct human connections (e.g., calling the former head of the National Cancer Institute).

Accelerating Patient Comfort with AI as a Healthcare Provider

A significant trend highlighted is the rapid increase in patient comfort with AI. A Bain survey (March 2024 - September 2025, 500 people) revealed that the percentage of consumers comfortable with "AI becomes my doctor" jumped from 11% to almost 30% in just one year, representing a 2.5x increase. This rapid pace of change, potentially reaching 50% comfort in another year, is unprecedented for the typically slow-moving healthcare system.

Drivers of AI Adoption: Headwinds and Tailwinds

The adoption of AI in healthcare is propelled by both "headwinds" (push factors) and "tailwinds" (pull factors):

• Headwinds: Increasing patient frustration with care due to physician and clinician shortages, limited time during patient visits, significant access issues (long wait times for appointments), and a growing lack of personalization where patients no longer "know their doctor." These factors increase the "substitutability" of care sources.
• Tailwinds: The convenience and ability of AI to address delays. Millions of users, both physicians and non-physicians, are submitting clinical queries to AI applications like OpenAI. This widespread usage is increasing the acceptance of information derived from non-human sources.

A Vision for AI-Augmented Healthcare: Beyond "Doctors Gone"

Dr. Mohiuddin argues against the "doctors are going to be gone" narrative, instead envisioning AI as a tool to address the "massive volume of care that isn't being done." AI can enable more effective care management, facilitate preventive care, and provide immediate answers to patient questions, rather than weeks later. The ultimate vision includes individualized LLMs (Large Language Models) or chatbots for every person, acting as a "quarterback of care" by knowing all their data and assisting with health management.

In this model, the physician's role evolves:

1. Expanding Panel Size: AI agents act as a "sidekick," managing care, handling administrative tasks, and addressing lower-risk health questions, thereby allowing physicians to expand their patient panels.
2. Focusing on Complex Cases: Physicians can dedicate more time to harder cases, leveraging their expertise and networks for specialized referrals, operating at the "top of their license."

This shift aims to provide end-to-end care and achieve value in ways not seen in the 15 years since the launch of CMMI and the push for value-based care.

Shifting from "Sick Care" to Proactive and Preventive Health

The current healthcare system often focuses on "sick care" because that's what is billable. AI, however, can enable a transition to proactive and preventive care. Examples include personalized nutrition counseling for conditions like diabetes, and individualized exercise plans that go beyond general guidelines (e.g., American Heart Association recommendations) to match personal preferences and metabolic states. This democratizes access to personalized health management, extending "smart board-certified help" to individuals who might not afford personal coaches or trainers.

Policy, Regulation, and the Natural Pull of AI Adoption

The discussion explores whether policy changes are necessary for AI adoption or if natural incentives suffice. Dr. Mohiuddin suggests that while policy can accelerate priorities and manage risk, over-regulation is a significant concern. The current US administration's approach is to learn, align on priorities, and foster a "coalition of the willing" (e.g., CMS pledges) to drive responsible adoption.

However, significant regulatory gaps exist. The FDA currently lacks mechanisms for post-deployment surveillance and monitoring for generative AI applications, especially for in-body devices. The proposed solution involves "experimental sandboxes" where government and industry entities collaborate on data sharing and developing appropriate regulatory frameworks.

Speaker 1 argues that, unlike previous digital therapeutics that required payment mechanisms, AI tools have a natural "pull" due to perceived value and ROI for health systems, driving adoption even without sweeping policy changes.

Current AI Adoption Trends and Critiques of Spending Metrics

A Menlo Ventures report on the "state of health AI" indicates high adoption in scribe and coding/billing applications, primarily by health systems, with less reported spend on the payer side. This suggests technology adoption is occurring organically.

However, Dr. Mohiuddin and Speaker 3 critique current spending charts as potentially misleading:

• Administrative Focus: Current adoption is heavily concentrated in administrative areas like SG&A (Selling, General, and Administrative expenses), OpEx (Operating Expenses), and revenue optimization (e.g., coding optimization). AI agents are replacing human tasks and coordinating complex administrative processes, representing only the "tip of the iceberg" for efficiency gains without additional regulation.
• Ambient Scribes: While seemingly clinical, ambient scribes also significantly benefit administrative applications by improving clinical documentation fidelity, which impacts revenue and care management.
• Limitations of Spending as a Metric: Speaker 3 argues that spending charts may become a poor way to track AI's impact due to the rapid reduction in costs, with intelligence "asymptotically heading to free." He predicts that administrative spend might decrease over time, with investment shifting towards more challenging areas like precision medicine and clinical trial matching.
• Missing Categories: Dr. Mohiuddin adds that current reports miss entire categories, particularly clinical innovation, which he expects to be "massive in three to five years." This includes improving provider productivity beyond ambient scribes (e.g., in-basket management, back-office operations) and leveraging better data for scalable precision medicine.

The AI Arms Race and Healthcare Costs: A Deflationary Vision

The concern arises that an "AI arms race" between payers and providers could increase net healthcare costs. Dr. Mohiuddin, speaking from UnitedHealth Group's perspective, asserts that their incentives are to bring price down while maximizing the impact on care delivery. He views UnitedHealth Group and its Optum divisions as committed to being a "deflationary force" in the industry. He acknowledges that other industry players have different incentives, leading to potential "increasing tension between payers and providers" before collaboration becomes the norm, requiring regulators to potentially intervene.

The Empowered Patient and Physician: A Bottoms-Up Disruption

A crucial "third or fourth trend" identified is the empowered patient voice, leveraging AI technology to influence traditional healthcare incentives. Similarly, empowered physicians are using AI (e.g., ambient technology) to make better decisions and potentially even operate more independently. This "bottoms-up" force, driven by patients and physicians, is seen as a new and powerful player that will fundamentally impact the healthcare landscape.

Conclusion: A Drastically Different Future for Healthcare

Dr. Mohiuddin's "hot take" is that the healthcare industry will look "drastically different in five years." He challenges incumbents to disrupt themselves, or risk a "significantly diminished role," as patients and citizens will ultimately drive the change by choosing options that serve them best. Unlike past technological pushes (e.g., high-tech and meaningful use, which often led to less satisfaction for doctors and patients), AI is different because it offers "meaningfully better" outcomes for the end-user. The ability of AI to execute "intelligent, highly intelligent tasks in a coordinated and orchestrated way," acting as a "boots on the ground workforce that is not human," will materially transform how care is delivered and potentially "who does it." This change, driven from the "bottoms up," is expected to be rapid and profound, fundamentally reshaping the healthcare ecosystem.