Key Takeaways

  • Americans over 60 reported $7.7 billion in scam losses during 2025, while AI-enabled schemes accounted for $893 million across more than 22,000 complaints.
  • Generative AI is making established fraud tactics cheaper, more convincing, easier to personalize, and practical to operate across many victims simultaneously.
  • Banks, telecom providers, advertising platforms, payment services, and government agencies face growing pressure to improve verification and accelerate intervention.

The Senate Special Committee on Aging heard a blunt assessment on Wednesday: Artificial intelligence is giving familiar fraud schemes industrial scale while erasing many of the warning signs consumers once relied upon.

The emotional levers remain familiar. Criminals still exploit fear, love, shame, urgency, and perceived authority. What has changed is their capacity to fabricate the evidence that makes those emotions persuasive, including a relative’s voice, a physician’s face, a bank’s telephone number, or an official-looking government seal.

A 64-year-old California woman described receiving a call in May from a man who claimed to have kidnapped her daughter. A crying woman whose voice sounded exactly like her daughter was placed on the line. The caller then threatened violence and controlled the victim's movements by telephone for 5.5 hours while she and her husband traveled among stores in the San Francisco Bay Area to transfer money.

After the caller claimed her daughter had been released at a Walmart, the woman contacted her directly. Her daughter was safely at work. The couple had already sent $5,400.

“It was the worst day of my life,” she told the committee.

The committee chairman, a Republican from Florida, said Americans over 60 reported losing $7.7 billion to scams in 2025. A bipartisan committee report, Artificial Intelligence & Older Americans: Confronting New Threats, Unlocking New Opportunities, attributed $893 million in losses to AI-enabled scams across more than 22,000 complaints.

Those figures illustrate why this is no longer solely a consumer-awareness issue. It is becoming an identity, payments, platform governance, and enterprise risk problem.

The American Bankers Association’s executive vice president for risk, fraud, and cybersecurity described generative AI as “industrializing” conventional scams rather than replacing them. Criminal networks can combine public records, breached accounts, social media posts, obituaries, and data broker records to construct highly personalized scripts. AI can also translate messages, remove grammatical errors, and help operators engage many targets at once.

A polished message is no longer evidence of legitimacy, nor is a familiar voice.

Banks are using analytics to identify unusual transactions, but the executive emphasized the continuing importance of employee judgment. A customer who appears frightened, secretive, or coached over the telephone may present signals that a transaction-monitoring model does not fully capture. That places branch employees, call-center staff, and payment operations teams on the front line.

The founder and medical director of the Roxbury Institute in Los Angeles detailed a parallel risk involving professional impersonation. Scammers combined footage from his YouTube channel with an AI-generated likeness and voice of a colleague, celebrity images, and copied media logos to advertise a supposed “miracle” lipedema cream.

The campaign lasted 11 days, despite reports to Meta, domain registrars, and search engines, and later resurfaced. Beyond direct financial losses, fabricated medical endorsements can lead patients toward unproven products, delay legitimate treatment, and damage trust in physicians whose identities were appropriated.

A representative from the law firm Crowell & Moring said deepfake-enabled fraud losses exceeded $200 million during the first quarter of 2025, citing an industry report. Deloitte has projected that generative AI could help drive U.S. fraud losses to $40 billion by 2027.

Can detection technology keep pace with content that can be regenerated almost instantly? The representative advocated wider use of media provenance systems, stronger coordination, and broader digital literacy. He also offered a low-tech safeguard: Families can establish private emergency code words and independently call a known number before sending money.

The Consumer Federation of America’s director of AI and privacy framed the threat as a wider “scam stack.” Data brokers help select victims, telecom and advertising systems deliver impersonations, and payment services move the proceeds. Fragmented reporting then slows recovery.

The FBI Internet Crime Complaint Center provides a central reporting channel, but the director said only 642 of more than 201,000 elder fraud complaints in 2025 entered the FBI Recovery Asset Team’s freeze process. Less than 0.5% of the reported $7.7 billion was frozen.

Proposals discussed at the hearing included a single ReportScams.gov portal, expanded cryptocurrency tracing, a national anti-scam strategy, stronger telephone customer verification, advertiser screening, and faster handling of impersonation complaints. The business implication is fairly direct: Fraud prevention increasingly depends on coordinated controls across the entire transaction journey, not another warning banner placed in front of the victim.