AI đang thổi bùng các vụ lừa đảo tiền bạc – đây là những gì bạn có thể làm để tự bảo vệ mình

The phone rings, and it’s your grandson’s shaken voice. There’s been an accident, he says, and he needs money before anyone finds out. Except it isn’t him. It’s software that learned his voice from a clip posted online, run by a stranger working through a list of phone numbers.

For years, warnings about artificial intelligence and cybersecurity have focused on corporate networks and government systems. Those threats are real. But if you follow the money in the FBI’s fraud data, a different picture emerges: Staggering sums are flowing out of household accounts.

In its 2025 annual report, the FBI’s Internet Crime Complaint Center began tracking complaints with an AI connection for the first time. Americans filed more than 22,000 such cases and reported roughly US$893 million in losses. Investment fraud accounted for $632 million of that, and people over 60 accounted for $352 million of the losses.

Those figures include only victims who reported to the FBI, and only cases where AI’s role could be identified. The consulting firm Deloitte projects that the actual hit from AI will be much bigger and help push U.S. fraud losses overall to $40 billion by 2027, up from $12.3 billion in 2023.

I’m a finance professor who studies household finance and how people use AI to make money decisions. And I believe the most consequential AI security story right now isn’t unfolding in server rooms, but at kitchen tables.

Old cons, new machinery

None of these scams are new. What AI changed is the cost and the quality of the cons.

Cloning a voice now takes a few seconds of audio and cheap consumer tools. In one study, listeners were able to identify an AI-generated voice only about 60% of the time. Video is heading the same way. In 2024, a finance employee at the architecture and design firm Arup was tricked into wiring about $25 million to fraudsters after they set up a video meeting “staffed” by deepfakes of the chief financial officer and several colleagues.

Phishing has improved, too. Clumsy wording and odd formatting used to give fraudulent emails away. Language models now write clean, fluent messages and can personalize them at scale using details scraped from social media. Deepfake videos of well–known business figures pitch bogus trading platforms.

The same pressure is visible in business losses. The cyber insurer Resilience reported that more than 85% of the losses in its claims portfolio in the first half of 2026 stemmed from attacks aimed at people rather than systems.

Why careful people fall for it

We’d like to believe that only careless people get taken. Research in behavioral finance says otherwise.

These scams are engineered around fear and urgency: a panicked grandchild, a boss demanding a same–day transfer, an investment window that closes tonight. Stress narrows attention and pushes people toward fast, intuitive judgments at the very moment they need slow, deliberate ones. Fraudsters also strike a pose of authority, whether a CFO’s face or a government agency’s letterhead, because most people defer to it.

Fluency matters as well. My own research examines how the smoothness of AI-generated communication leads people to trust it. A message with no typos, in a voice that sounds exactly right, sails past defenses that a clumsy fake would have tripped.

Nobody plans to make a major financial decision mid–panic. That’s exactly why scammers manufacture the panic.

The quiet version of the attack

AI-enabled theft doesn’t necessarily involve talking to the victim. Stolen personal data sells for a few dollars on dark web markets. Criminals then feed it to automated AI agents that probe bank and fintech systems around the clock, testing credentials and hunting for weak points at a speed no human crew could match. Last fall, the AI company Anthropic disrupted an espionage campaign in which an AI agent performed 80% to 90% of the intrusion work against roughly 30 targets, including financial institutions.

When an attacker gets into a customer account, the takeover can be over in minutes. Instant payment platforms like Zelle, built for speed and convenience, become the getaway car. The money typically moves within minutes, and getting it back is almost impossible.

How to protect you and your loved ones

Banks defend their own wire rooms with procedures, not vigilance. Households can borrow those procedures.

Check by calling back. When you get a suspicious call, hang up and dial a number you already know – like your bank’s fraud hotline – and never one the caller or message supplies. The point is to leave the channel the scammer controls. A cloned voice can’t answer your grandson’s real phone.

Trust and verify. You should agree on a robust family code word for emergencies and treat any request for money that lacks it as fake. Require two people in your household to sign off on any large transfer, so nobody moves serious money alone and under pressure. And build in a delay, such as a self–imposed 24-hour wait, before any big payment. Urgency is the scammer’s tool. Slowness is yours.

Add layers of protection. Protect the accounts themselves, too. Turn on two–factor login for financial accounts, and never share a verification code with someone who contacts you. That code is the second lock on your door, and the only reason a caller wants it is to get in.

You should also switch on your bank’s transaction alerts so a takeover announces itself in minutes, and consider a credit freeze, which is free and blocks thieves from opening new accounts with data bought on the dark web.

Protect older people. With $352 million of reported AI-related losses coming from Americans over 60, conversations with older family members are key. Walk through the callback rule and the code word, and ask their bank or brokerage about adding a trusted contact they can call. It costs nothing, and it gives the institution a way to raise an alarm before the money moves.

If money has already moved, you should call your bank immediately, ask it to attempt a recovery, then report the scam to the Federal Trade Commission.

Where habits end, rules should begin

Good habits raise the cost of every one of these scams, but they can’t do it all. This is where the U.S. regulations have fallen behind the technology.

Federal law is supposed to protect consumers from unauthorized electronic transfers, and regulators have said that a transfer set in motion by a fraudster counts as unauthorized even when the victim was tricked into handing over account credentials.

In practice, though, victims of instant–payment fraud often recover little. Banks frequently classify losses as “authorized” when a customer was deceived into approving the payment, and even obvious victims of such takeovers can face long fights over reimbursement.

For example, the Consumer Financial Protection Bureau sued Zelle’s operator and three of the country’s largest banks over their response to alleged fraud in late 2024, then dropped the case in March 2025. New York’s attorney general has since filed her own lawsuit, which a judge allowed to proceed in July. Zelle’s operator denies the allegations and says it will appeal.

The U.K. has taken a different path. Since late 2024, U.K. banks have been required to reimburse most scam victims up to £85,000, roughly $115,000, with the cost split between the sending and receiving institutions. The logic is that banks are best equipped to fight fraud, since they run security teams and networkwide analytics that can spot suspicious patterns across millions of transactions. What they had lacked was a strong financial reason to deploy them fully, and the reimbursement rule supplied it.

The regulator’s own dashboard shows that 88% of the money lost to eligible scams has been returned to victims since the rules took effect. More telling: An independent evaluation found that scam losses fell by roughly a fifth in the rule’s first year. That shows that when banks bear the losses, they find ways to prevent them.

I believe American regulators and Congress should study that model closely. When payments are instant and irreversible, the risk cannot rest almost entirely on the customer, who is the least–equipped party in the chain.

Pawan Jain does not work for, consult, own shares in or receive funding from any company or organization that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.