AI telehealth risks and the MEDVi moment.

Our industry has been captivated by the idea of AI-driven scale—fueled in part by what some have called the "Sam Altman prophecy." The notion that a one-person, billion-dollar company could emerge through AI isn’t just theoretical anymore. It’s being tested in real time.

On the surface, the numbers behind MEDVi were compelling. $401 million in sales in 2025, with projections nearing $1.8 billion this year. It was positioned as a breakthrough, described as a benchmark for modern AI marketing.

But beneath that growth lies a more important question: what are the real AI telehealth risks when speed outpaces trust?
MEDVi attracted over 500,000 users and widespread attention online. Many believed they were interacting with licensed clinicians. In reality, much of that experience was powered by AI-generated personas.

The platform, which markets itself as “doctor-approved”, offers lower-cost alternatives to medications like Ozempic and Wegovy. But scrutiny from the U.S. Food and Drug Administration revealed the sale of unapproved products—raising regulatory concerns and triggering enforcement actions in early 2026.

This moment isn’t just about one company. It’s a signal.

AI telehealth risks vs. the human foundations of care

The rise of AI in healthcare is not inherently problematic. In fact, I think its potential is significant.
Many individuals already turn to AI tools for health-related guidance. In certain contexts, these systems can support key elements of care:

• Time: Always available
• Transparency: A space for open, judgment-free questions
• Transition: Organizing complex health information
• Trust: When used appropriately, helping people feel informed

But these benefits depend on one critical condition: honest design.
When AI is positioned as a support system, it can extend care. When it is positioned as a replacement, especially without disclosure, it introduces risk.

In cases like MEDVi, where AI was presented with a human face, the line between assistance and deception becomes blurred. And when that happens, trust erodes quickly.

The age of AI telehealth: where do we go from here?

AI-powered telehealth is often framed as a solution to cost and access challenges. And in many ways, it can be.
Healthcare is not just the delivery of information. It is the interpretation of nuance, the ability to intervene, and the responsibility to act in real time. These are not easily replicated.

As the Gartner Hype Cycle suggests, we’re currently navigating the “Trough of Disillusionment” where we’re looking in hindsight at what went wrong with MEDVi’s marketing hype. I hope we reach the “Slope of Enlightenment”, salvaging what worked and what we could do to ensure the next telehealth platform meets ethical standards.

Deepfakes and the regulatory investigations aren’t just symptoms of a bad business model. They are symptoms of a design that forgot the human, the person behind the screen, for the sake of innovation and the bottom line.

We don’t need more ‘pied pipers’ in patient care. We need intentional design that doesn’t just say ‘we care’ but actually demonstrates it according to these Rules of Medicine.

Get insights like this delivered monthly.

Subscribe to the CancerGeek newsletter for exclusive perspectives on healthcare innovation, digital health trends, and more delivered straight to your inbox.