La FTC informó que en 2025 hubo más de un millón de reportes de estafas de suplantación y $3.5 mil millones en pérdidas reportadas. El FBI también ha documentado campañas que utilizan voces generadas por IA para hacerse pasar por personas reales.
The old scam was easy to recognize: a badly written email from a stranger asking for money. The new one may know your name, your bank, where you live—and perhaps even sound like someone you love. Artificial intelligence is making fraud more personal, more convincing and much harder to recognize.
According to a new Consumer Reports survey, the chances are overwhelming that you have – a whopping 91% of Americans reported they’ve encountered a digital scam attempt like scam messages, phishing texts, fake invoices, and scammers posing as a bank or person they know. Seventeen percent have lost money to a scam. Behind these alarming stats are the stories of real people. One shopper told Consumer Report they spent weeks watching what looked like the perfect deal on Facebook. The ad used a real manufacturer’s name, logo, and address. He ordered, but the product never came. Another shopper bought memory supplements after seeing what she thought were celebrity endorsements. They turned out to be AI-generated deepfakes.
The message arrives on your phone. “Dad, I need your help.” It appears to come from your daughter. Perhaps the number is unfamiliar, but the explanation comes immediately: she lost her phone. There has been an accident. She is traveling. Something happened to her bank account. She needs money quickly. Then the phone rings. The voice sounds like hers. Now imagine that the caller knows her name, knows yours and mentions something about the family that a stranger shouldn’t know. A few years ago, those details might have been enough to convince you that the call was real. Today, they may not prove anything. Welcome to the new age of scams.
The Bad Grammar Was Protecting Us
Many traditional online scams practically announced themselves. The email contained strange capitalization. The grammar was terrible. The logo looked wrong. A mysterious millionaire wanted your assistance moving money out of another country. Someone you had never met had selected you for an improbable inheritance. The absurdity was almost reassuring. We learned the warning signs. But artificial intelligence is eliminating many of them.
A scammer no longer needs to write convincing English. Generative AI can produce a polished business email in seconds. It can adjust the tone, translate the message into another language and rewrite it until it sounds professional. That matters because many of the clues consumers were taught to recognize are becoming less reliable. The suspicious message may no longer look suspicious.
A Scam Written Especially for You
The internet already contains an extraordinary amount of information about ordinary people.
Your social-media accounts may reveal where you work, where you vacation, the names of relatives, restaurants you visit, organizations you belong to and perhaps even when you are away from home. Professional websites can reveal your employer and colleagues. Public records and data breaches can add addresses, phone numbers and other information. A scammer doesn’t necessarily need all of it. Sometimes one or two accurate details are enough.
Imagine receiving this: “Hello, Robert. This is the fraud department. We’re calling about suspicious activity on your account.” Generic.
Now imagine: “Hello, Robert. This is the fraud department at your bank. We’re calling about suspicious activity involving your checking account.” If the bank named in the message happens to be your bank, the psychology changes immediately.
The FTC says scammers use texts, calls, emails, social media, search results and other channels to impersonate businesses and government agencies. In 2025, consumers reported losing almost $1 billion to business impersonators alone, with bank impersonation producing the highest reported losses within that category. Federal Trade Commission
The scam works because it begins with something true. Then comes the lie.
Now the Voice Can Be Fake Too
For generations, hearing someone’s voice provided a kind of authentication. You might distrust an email supposedly sent by your son. But if you heard his voice on the telephone, crying and asking for help, skepticism became much harder. AI is weakening that assumption.
Voice-cloning technology can generate speech resembling a real person. The FBI has already warned about malicious campaigns using AI-generated voice messages to impersonate real people. FBI that creates a frighteningly effective combination: publicly available information supplies the story, while artificial intelligence can help provide the voice.
Imagine receiving a call late at night. “Grandma, please don’t tell Mom. I’ve been arrested.” Or: “Dad, I’ve been in an accident.” The voice sounds familiar. The caller is frightened. Another person comes on the line claiming to be a police officer, lawyer or doctor. Money is needed immediately. The scammer’s most important weapon is not the technology. It is urgency. Because urgency prevents verification.
Fear Is the Real Technology
The most successful scams often create an emotional emergency before they ask for money. Your bank account has been compromised. You owe the government money. Your computer has been hacked. Your grandson has been arrested. A package cannot be delivered. Your toll payment is overdue. Someone is using your identity.
Each story is different, but the mechanism is remarkably similar: something terrible is happening, and you must act now.
The FTC reported that government impersonation scams increased sharply in 2025, helped in part by bogus messages about unpaid road tolls. Those messages sometimes imitate actual toll systems and threaten additional fees or suspension of vehicle registration if the recipient does not pay immediately. Consumer Advice
A sophisticated scam therefore does not have to convince you forever. It only has to convince you for five minutes. Long enough to click the link. Long enough to reveal a password. Long enough to transfer the money.
Your Bank Will Never Need You to “Protect” Your Money From Your Bank
One of the most dangerous variations begins with a supposed security warning. Someone claiming to represent your bank, a government agency or a technology company tells you that criminals are attempting to steal your money. Fortunately, the caller says, they can help. You simply need to move your savings somewhere “safe.” That safe account belongs to the criminals.
The FTC says some of the costliest impersonation scams begin with fake security alerts and persuade victims to move money supposedly to protect it. The amount lost can be enormous because the scam is specifically designed to make the victim believe that not transferring the money is the dangerous choice. Federal Trade Commission
This suggests one extremely useful rule: Anyone who unexpectedly contacts you and tells you that you must move money to protect it should immediately make you suspicious. Do not continue the transaction simply because the caller knows personal information about you. And don’t trust the telephone number displayed on your screen as proof of identity. End the conversation. Then contact the institution yourself using a telephone number or website you independently know to be legitimate.
Create a Family Password
There is another simple defense against AI voice impersonation, and it requires almost no technology. Create a private family verification word or phrase. If someone calls claiming to be your child, grandchild, spouse or another relative and something feels wrong, ask for it. Better still, don’t depend on the incoming call at all. Hang up and call the person directly using the number already stored in your contacts. If you cannot reach them, contact another family member who can verify what is happening. The principle is more important than the particular method:
Use a second channel to verify the first.
If the message comes by email, call. If the request comes by text, contact the person another way. If someone calls claiming to represent your bank, hang up and call the number on the back of your card. The scammer wants to keep you inside the story. Your job is to step outside it.
Social Media Gives Scammers a Window Into Our Lives
There is an uncomfortable irony in all of this. We have spent years voluntarily publishing information that can make impersonation easier. A birthday photograph identifies family members. A vacation post reveals where we are. LinkedIn shows our professional relationships. Facebook can reveal relatives and friends. Instagram provides photographs and video. Other public databases fill in additional pieces. None of that means people should abandon social media. But it does mean we should understand that information posted for friends may also be visible—or eventually become available—to strangers.
The FTC reported that nearly 30 percent of people who reported losing money to scams in 2025 said the scam began on social media, with reported losses reaching $2.1 billion. The agency notes that scammers can exploit what people post to determine how to target them. Federal Trade Commission
AI makes assembling those fragments easier. What once required time and research can increasingly be organized and transformed into a convincing message much faster.
The Five-Minute Rule
Perhaps the most effective security tool isn’t an antivirus program. It is time. When someone unexpectedly contacts you demanding money, information or immediate action, deliberately interrupt the urgency. Give yourself five minutes. Do nothing. Do not click the link. Do not provide the verification code. Do not transfer the money. Do not allow the caller to convince you that hanging up will make the emergency worse. Then investigate independently. Scammers manufacture urgency because calm people ask questions. And questions are dangerous to a scam.
Trust Has Changed
For years, cybersecurity advice focused heavily on technology: use strong passwords, update your software, install security tools and avoid suspicious links. Those things still matter. But AI is pushing security into a different territory. Increasingly, we also have to protect something much older than our computers: our trust.
A familiar voice may not prove identity. A professional email may not prove legitimacy. A photograph may not prove an event occurred. A caller knowing your name may not mean the caller knows you. Even an accurate personal detail may simply be the bait that makes the next false statement believable. That doesn’t mean living in permanent suspicion. It means developing a new habit: when an unexpected request involves money, passwords, personal information or urgency, verify before you act. The scams of the past often succeeded because people knew too little. The new scams may succeed because the scammer appears to know so much. And the most dangerous message may no longer begin: “Dear Sir or Madam.” It may begin with your name.
There’s a bipartisan bill in Congress called the SCAM Act that would require online platforms to verify who’s behind a paid ad, investigate reported scams fast, and take fraudulent ads down. You can also read the new Cyber Readiness Report of Consumer Report to see who’s most likely to lose money to a scam, what finally makes people hang up or click away, and how you can protect yourself.
