On April 2nd, 2026, the New York Times published what the internet immediately declared the greatest entrepreneurship story of the decade: a self-taught founder in Los Angeles built a $1.8 billion telehealth company with two employees, $20,000, and a stack of AI tools. Sam Altman said he wanted to meet the guy. LinkedIn exploded. Inc. Magazine ran a follow-up. There was just one problem — the story the Times told left out the 800 fake doctors, the AI-generated patient photos, and the FDA warning letter that had arrived six weeks earlier.
The Story the World Was Told
Matthew Gallagher, the founder of a telehealth company called Medv, built a platform selling compounded GLP-1 weight loss drugs online — the same active ingredients as Ozempic and Wegovy, but cheaper. Customers sign up, complete an online consultation, and the drugs ship to their door. The financials, which the Times verified, were staggering: $41 million in revenue in its first full year, a 16.2% profit margin, 250,000 customers, and a projection of $1.8 billion for the following year. The AI tools Gallagher used — ChatGPT, Claude, Grok, MidJourney, Runway — were held up as proof that the future of business had arrived. Every entrepreneur on the internet agreed.
Anyone with experience in direct response marketing and FTC compliance, however, would immediately ask one question: how does a two-person company scale to those numbers in a regulated health category without a compliance team?
What Was Actually Happening
800 Fake Doctors on Facebook
The actual customer acquisition engine behind Medv was not a brilliant AI-powered marketing machine. Gallagher and his affiliates created over 800 Facebook profiles impersonating doctors — entirely fabricated human beings with AI-generated headshots, professional credentials, and thousands of active ad campaigns promoting Medv's weight loss drugs.
08:45
Screenshots of fake doctor Facebook profiles including 'Dr. Alistair Whitmore' and 'Professor Albus Dongor'
Watch at 08:45 →
When investigators began examining these profiles, the details collapsed immediately. One profile for a supposed weight loss physician listed the doctor's location as Kiev, identified them as female, and listed their interests as Bronze Age Asian archaeology. Another fake physician had an Angolan phone number. A third had contact information linking to a clothing store in the Republic of Congo. The most brazen profile was listed simply as Professor Albus Dongor. At peak operation, there were 5,000 active ad campaigns on Facebook running under these fake doctor personas, each selling weight loss drugs to real people who believed they were receiving guidance from a licensed physician.
AI-Generated Before-and-After Photos
The Medv website was populated with dramatic patient transformation stories. Melissa C. lost 40 pounds. Sandra K. lost 35. Michael P. lost 50. None of these patients existed. Futurism investigated and confirmed the before-and-after photos were AI-generated deepfakes — the weight loss never happened, the testimonials were fabricated. When Futurism published its findings, the images were quietly swapped out, but the names remained. The same names — Melissa C., Sandra K. — were now attached to entirely new AI-generated faces. In one replacement image, Sandra K.'s fingers are visibly melted together on her smartphone, a telltale artifact of AI image generation.
12:30
AI-generated before-and-after patient photo showing fused fingers on smartphone — a common AI image artifact
Watch at 12:30 →
The FDA Warning Letter Nobody Mentioned
On February 20th, 2026 — six weeks before the Times profile ran — the FDA sent Medv a formal warning letter (Warning Letter No. 721455). The agency cited multiple violations: the site falsely implied Medv compounded the drugs itself, when in fact it is a marketing company that routes customers to third-party platforms called Care Validate and Openloop. The FDA also flagged language claiming the drugs shared the same active ingredient as Wegovy and Ozempic, noting that such language falsely implies FDA approval for compounded drugs that have not been evaluated or approved. The warning letter explicitly noted that the violations cited were "not intended to be all-inclusive" — meaning regulators believed there were likely more problems than the ones identified. The Times did not mention any of this.
One month before the Times profile, Medv's clinical partner Openloop suffered a data breach in which a hacker claimed to have stolen records from 1.6 million patients, including names, addresses, dates of birth, and medical treatment history. Half a million people had signed up through fake doctor profiles, seen fabricated before-and-after photos, purchased drugs the FDA had flagged, and now had their medical records exposed.
A 100-Year-Old Playbook, Supercharged by AI
The technique Medv used — manufacturing fake expert authority to sell a product — is not new. It was formalized by Edward Bernays in the 1920s. Bernays understood that consumers don't trust companies, ads, or salespeople, but they do trust doctors, scientists, and independent authorities. His solution: don't sell the product yourself; get a credible expert to sell it for you.
When the pork industry needed to increase bacon sales, Bernays didn't run ads. He recruited real physicians to publicly endorse heavier breakfasts as healthier for Americans, then tied that endorsement to bacon and eggs. Medical recommendation replaced marketing copy, and sales surged. When American Tobacco needed to sell more cigarettes to women, Bernays arranged for real doctors to endorse smoking as an aid to weight management. Real doctors, wrong advice, but real people making real statements.
Gallagher took Bernays's framework and ran it through AI. Instead of spending months cultivating relationships with real physicians, he generated 800 fabricated ones in an afternoon. Same psychological mechanism — manufactured third-party authority — but AI made it infinitely scalable and nearly free. The manipulation is identical. The doctors are not.
The Compliance Reality and What Comes Next
The National Consumers League has already filed a formal request for an FTC investigation, stating that Medv's conduct violates the FTC Act. The FTC's specific areas of concern align precisely with Medv's business model: deceptive advertising through fake endorsers, AI-generated testimonials presenting non-existent results, and false equivalence claims regarding FDA-approved medications. Every pillar of the company's customer acquisition strategy involves conduct the FTC and FDA have explicitly prohibited.
Legitimate marketers bear a hidden cost when stories like this emerge. Regulators tighten rules for every advertiser. Platforms restrict ad policies across the board. Consumers grow more skeptical of everyone in the category. One operation that ignores compliance raises the cost and difficulty of operating ethically for every competitor who doesn't.
For any entrepreneur building in regulated categories — health, supplements, financial products — the compliance framework is not something to address later. The FTC pursues companies doing $10,000 a month with the same authority it uses against billion-dollar operations. The rules do not scale with revenue.
Building Authority Without Building a Fraud
Authority is a legitimate and powerful marketing tool. Bernays was correct that people buy from those they trust, and trust flows toward expertise. The goal is not to abandon authority-building — it is to build it on a foundation that holds up to scrutiny.
- Use real testimonials from real customers. If your product works, customers will describe the results. Ask them directly, make documentation easy, and record their actual stories. A single verified testimonial from a genuine customer carries more persuasive weight than a thousand fabricated ones and will not generate regulatory action.
- Build real relationships with real experts. Find a credentialed professional who genuinely believes in your product. Pay them fairly, give them editorial control, and let them speak in their own voice. One authentic endorsement from a real expert is legally defensible and more credible than any number of AI-generated stand-ins.
- Know the regulatory rules before you run your first ad. In health, supplements, and weight loss categories, the FTC and FDA govern what claims you can make, how testimonials must be disclosed, and what language implies approval you do not have. This is not optional and not something to address at scale — it applies from day one.
- If you need fake experts, examine the product. Marketing that only functions when endorsed by people who do not exist is a signal about product quality, not a marketing problem. Fix the product. Real results generate real testimonials and real expert interest.
The detail that captures the entire story is Sandra K.'s melted fingers. Someone at Medv or one of its affiliates looked at that image — fingers fused to a smartphone in a way no human hand produces — and approved it for publication. They ran it. That is what happens when the people you are selling to stop being real to you. You stop noticing when they are not.
A business built on fabricated doctors, fabricated patients, and fabricated results is not a glimpse of the future of business. It is a glimpse of what the tools now make possible for anyone willing to use them this way. The technology does not make the choice. That part has not changed.
