How AI is changing the threat for families in 2026, and what to do about it.
I have been helping large, global companies manage their documents and information going on the better part of three decades. I’ve founded three companies related to managing documents with my last one focused on real estate transactions. That company, Cartavi, was acquired by Docusign and I’m proud to see it still a part of their offerings thirteen years later. I had a senior role at Docusign for nearly years after the acquisition and helped them own the real estate space as well as move beyond eSignatures. I have spent more time than I want to admit thinking about how people share, store, and protect the documents that matter most to them.
What I am seeing in 2026 does not look like the fraud landscape I knew even just five years ago. The numbers tell part of the story. Americans lost nearly $21 billion to cybercrime and scams in 2025. Adults over 60 alone lost approximately $7.7 billion, more than any other age group. Over 211,000 Americans over 60 filed fraud complaints with the FBI last year. The average reported loss for older fraud victims exceeded $38,000, and more than 12,000 older adults lost over $100,000 each. Fraud targeting older adults has quadrupled since 2020.
Those numbers are bad. The bigger problem is what is driving them.
AI is changing the math
Five years ago, a fraudster trying to impersonate someone’s daughter needed to know a few personal details, sound vaguely convincing on a phone, and hope the person on the other end was confused enough to send money. The scam worked sometimes. It also failed often, because the human running it was juggling dozens of calls and the script was easy to spot.
That is not how this works anymore. AI tools available today can clone a voice from 30 seconds of audio. They can generate video that looks photographically real. They can write personalized phishing emails that reference your actual relationships, your actual employer, your actual recent activity. The scam is no longer a person on the other end of a phone. It is an automated system that gets more convincing with every interaction.
The result is a category of fraud that did not exist at scale until recently. Romance scams using AI-generated images and voice. Tech support scams using deepfake video of trusted faces. Family emergency scams where a "grandchild" calls in real distress, in their actual voice, asking for bail money or a wire transfer. Investment scams that show you a real-time dashboard of fictional gains.
Every one of these is harder to spot than its pre-AI predecessor. And every one of them is being run at higher volume than was possible before.
The five faces of modern fraud
From our work with families and the patterns I see across our partner organizations, some of the most active fraud categories targeting families in 2026 break down into five broad shapes.
The first is the impersonation scam:
Someone calls or messages claiming to be a family member, a doctor, a government official, a bank fraud department. The voice is convincing because it is. AI made it that way.
The second is the document scam:
A forged bill, a fake collection notice, a counterfeit IRS letter that looks exactly like the real thing. The goal is to get you to call a number, send a payment, or hand over enough information to enable the next step.
The third is the tech support scam:
A pop-up on your computer says you have a virus and to call a number for help. The "technician" on the other end is the fraudster. By the time you have given them remote access, the damage is done.
The fourth is the romance scam:
Someone builds a relationship with a target over weeks or months, then asks for money for an emergency. AI has made these dramatically more convincing and dramatically more common. Investment and cryptocurrency scams, often layered on top of a romance setup, generated more than $11 billion in losses in 2025.
The fifth is the disaster scam:
After a hurricane, a wildfire, a flood, fraudsters pose as FEMA, as insurance adjusters, as contractors. They target the people whose documents and proof of property are hardest to access at exactly the moment they need them most.
What I tell families
Three things, in order of importance.
Talk to your family about this now, while everyone is calm. Decide on a “family word” that you would use if anyone in the family ever calls asking for money or help. If the person on the phone cannot say it, the call is fake. This single practice would prevent more impersonation scams than any technology I am aware of.
Slow down before you respond to anything that creates urgency. Real banks, real government agencies, real family members do not pressure you to act in the next five minutes. The pressure itself is the tell.
Get your information out of the places where it lives unprotected. Email, sticky notes, paper files, a folder on a home computer, a password written on a desk pad. These are the surfaces where fraud starts. The point of a secure digital vault, the one we built and the one AARP made available to their members, is that the information lives in one place with controls you understand and access you decide.
The fraud landscape has gotten more sophisticated. The defenses are simpler than most people think. Most families just have not been told what to do.
That is what this work is about for me. The technology I have spent my career on is most valuable in the moments families need it most. Hurricane season is one of those moments. A scam call from a "grandchild" is another. The next one, whatever it is, is the one nobody saw coming.
Get the conversation started. The rest gets easier from there.