The Consumer Technology Association recently convened Health Week , a summit of health policymakers, digital consumer brands, academics, entrepreneurs and investors to debate the future of medicine and artificial intelligence.

Rene Quashie , Vice President of Digital Health at CTA, opened the event with a personal experience revealing how ChatGPT was instrumental, if not insistent, in helping him seek urgent medical attention. The tool later proved more helpful than his physician in helping him navigate a prolonged recovery. Quashie highlighted new CTA research showing 41 percent of consumers now incorporate AI and/or social media into their non-emergency healthcare journeys, with AI emerging as the preferred digital source. His remarks framing AI as a superior alternative to our “idealized version of healthcare” set the stage for panels like “Should AI Be Allowed to Practice Medicine?” and “The AI Divide: Who Benefits and Who Gets Left Behind?”

In the first debate, American Medical Association CEO John Whyte expressed discomfort with AI operating absent physicians and cited the inaccuracy of AI-based diagnostic tools like blood pressure cuffs in retail kiosks. Marc Paradis countered, “There is nothing magical about being human,” arguing that we unfairly measure AI against the world’s best doctor rather than the deficient care most people actually receive. Laura Adams , a National Academies of Sciences, Engineering, and Medicine (NASEM) expert, noted, “Economics will decide who gets to do this. What does it mean for the patient and where can the tool sit most easily in the patient’s path?” The discussions, while provocative, did address concerns about access. Will these game-changing technologies be democratized for everyone, regardless of zip code and income, or only a few?

Mistaking Necessity for Democratization

To support the ubiquity and utility of AI in healthcare, several speakers referred to data from Pew Research, Rock Health and the Kaiser Family Foundation showing that lower-income and non-White populations use chatbots and social media disproportionately for health information. The tone across these comments suggested we have reached an age in which these technologies are being democratized. Some in the room were surprised by data showing that some low-income households first turn to health chatbots for support and information.

My interpretation of the data is not a reason to celebrate. Lower-income communities may increasingly rely on AI-generated tools, but compared to what? Marginal care? No care? Unaffordable care? ​

What The Data Show About Chatbot Use

At times, aggregate data and summaries omit some of the most insightful research information. In general, Pew Research data show 34% of adults use AI chatbots for health. In contrast, KFF data reveal uninsured adults are more than twice as likely as the insured to use AI for mental health advice due to sheer affordability and appointment backlogs.

However, a more granular review of June 2026 Pew Research Center data shows that compared to lower-income adults, higher-income adults used AI chatbots for at least one health purpose almost twice as much, with 28% vs. 48% use among lower- and upper-income adults, respectively. Educational attainment was similar, with 23% vs. 44% among high school vs. four-year college graduates, respectively.

Rock Health’s December 2025 survey of 8,000 Census-matched adults showed parity in use across income and education brackets. However, both Pew and Rock Health reported higher adoption among non-White participants. Kaiser Family Foundation data from early 2026 show 32% of adults turned to AI for health information in the prior year due to affordability, no provider or inability to access appointments. Uninsured adults were more than twice as likely as insured adults to have used AI for mental health advice, 30% compared with 14%. My interviews with Medicaid enrollees confirm they use free chatbots out of necessity, not preference. Even then, they would prefer to corroborate the information with a human they trust.

Relying on AI for healthcare because you cannot afford or access care reflects our payment-driven system and may explain why Medicaid was barely mentioned at the meeting. Altogether, the use of AI tools and chatbots is a sobering reflection on healthcare access for many Americans. Thus, turning to AI tools like chatbots for health information and support should not be conflated with access to care, diagnostics and treatment needed to improve population health.

The meeting left me feeling as though we are barreling down the AI track with warnings to get on board or be left behind. But how will AI innovation be distinguished from past digital health advances that have yet to impactfully and sustainably reach those in greatest need?

Will better outcomes continue to accrue primarily to those with the most access? Will we continue shifting the burden to payment models and policymakers to explain our unrelenting disparities?

Quashie described the purpose of Health Week: “As AI and advanced technology become more ubiquitous in healthcare, focused convenings like CTA Health Week create opportunities for government leaders, tech companies, health systems, and clinicians to learn from one another and shape the future of healthcare together.”

If that mission means anything, then it must begin by confronting the systemic failures that make a chatbot a patient’s last resort.