AI-Powered Scammers Are Draining Medicare Billions, and Washington Can't Keep Up

A $1 billion catheter scam that never touched American soil

Criminal networks operating out of Russia, Eastern Europe and Asia are stealing billions of taxpayer dollars from Medicare, and they're doing it without ever setting foot in the United States, according to a New York Post investigation.

The mechanics are almost absurdly simple once you see them. Buy up dormant Medicare-enrolled supply companies. Keep the old ownership names and bank accounts on file so nothing looks changed. Use stolen patient data to file claims. Collect.

Prosecutors say one Russian-linked network bought 30 U.S. medical supply companies already enrolled as Medicare providers, then used them to bill Medicare roughly $10 billion, mostly for urinary catheters nobody ordered or needed. The scheme used stolen data from more than one million Americans and triggered 400,000 complaints, according to the indictment cited by the Post. Most of the fraudulent reimbursements were blocked, but the gang still laundered large sums through banks in Israel, Turkey and China, prosecutors charge.

A change in Medicare provider ownership is supposed to trigger a site visit. It didn't happen, because the network sent people from Russia and Estonia to physically pick up paper checks in the U.S. and never notified the Centers for Medicare and Medicaid Services that ownership had changed at all.

AI voice cloning is the new weapon of choice

Gordon Schnell, a partner at the healthcare fraud whistleblower firm Constantine Cannon, told the Post the sophistication of what he's seeing "blows me away." His point about scale matters: CMS distributes an enormous amount of money every year, so even a small percentage lost to fraud adds up to billions.

A March 2026 report from the cybersecurity firm Pindrop, submitted to CMS, laid out how this actually works. Fraudsters use cloned voices to authorize transactions at machine speed. Automated AI voice bots probe Interactive Voice Response systems, the same phone menus you get when calling any large company or government agency, testing for weaknesses before escalating to a live human agent. The goal is to trick that human into releasing account data.

According to the Post, one Philippines-based operation went further, training AI systems to mimic the speech patterns of elderly Americans specifically so they could fool insurance companies calling to verify suspicious claims. Synthetic identities built from breached data are also used to fake enrollment or manufacture consumer consent, per the Pindrop findings.

It's not just billing systems. It's grandma's phone.

CMS's own CyberGeek blog, published February 11, 2026, documented a real vishing call that hit a family member of a staff writer. A caller, polite and professional-sounding, told an elderly woman her new Medicare card was already in the mail. Over the course of an hour, she and her husband gave up their Medicare numbers, address, medication history and coverage details. The red flag only went up when the caller pushed a back brace nobody asked for.

When she hung up and the caller phoned back immediately, that confirmed something was wrong. CMS notes AI now lets scammers scrape personal data from public records, data brokers, social media and leaked databases in minutes, then stitch it into a believable script. Once a number is confirmed responsive, it gets shared across criminal networks, meaning victims typically see a spike in follow-up scam calls.

The SMP Hawaii nonprofit, which has worked to protect the state's elderly residents since 1997, documents the same toolbox: AI-generated false claims, fabricated patient identities, deepfake audio used to fake doctor-patient consent for prescriptions or procedures, and even altered medical imaging to justify fraudulent surgeries or claims.

The dollar figure nobody disputes

The New York StateWide Senior Action Council, which runs New York's federally funded Senior Medicare Patrol program, put a number on the broader problem in a September 2025 release: Medicare fraud costs taxpayers more than $60 billion nationally every year, according to executive director Maria Alvarez. That figure predates the specific AI-driven schemes now surfacing and likely undercounts the newest tactics.

Alvarez's advice to seniors is standard but sound: review every Medicare statement, be skeptical of unexpected calls, never click links in unsolicited texts or emails, and independently verify any medical provider before sharing information.

What CMS is actually up against

A fair-minded skeptic might ask whether this is overstated, since AI has legitimate uses in fraud detection too, and CMS does catch and block plenty of bad claims, including most of the $10 billion Russian-linked scheme. That's true and worth stating plainly. But the site-visit failure in the catheter case shows the verification process itself has a hole large enough to drive a truck, or in this case 30 shell companies, through.

No new legislation or CMS rule change has been announced in these reports to close the ownership-verification gap that let the catheter scheme run for as long as it did. Whether Congress or CMS moves to require faster, harder ownership-change verification, and whether that would even outrun AI-driven fraud that already operates faster than human reviewers can react, remains an open question.