$68 billion stolen, America faces the AI scam tsunami

Introduction: a bill that makes your head spin

A report that confirms the worst fears

A report published on June 30, 2026 by Gallup and the Stop Scams Alliance reveals that 15 million Americans were victims of scams in 2025, for total estimated losses of $68 billion, or roughly $186 million a day. That figure alone exceeds the combined annual budget of several U.S. federal agencies, and it lands at a moment when most households already feel squeezed by everyday costs, which makes the scale of this theft even harder to stomach.

The most troubling part of this report remains the growing role played by artificial intelligence: 12% of successful scams in 2025 involved AI tools or deepfakes, a share that would have seemed unthinkable just three years ago, back when most fraud still relied on clumsy emails and obviously fake phone scripts that a careful reader could usually spot.

A survey conducted on a large sample

The survey, conducted among 5,173 American adults between January 8 and February 18, 2026, reveals that 6% of American adults were scammed last year, and that 24% of them have already been victims of a scam at least once in their adult life.

This level of methodological detail, rare for this type of survey, gives the Gallup and Stop Scams Alliance report particular credibility, reinforcing the idea that these figures are not a rough estimate but a rigorous measurement of a rapidly expanding phenomenon that touches nearly every demographic group, from retirees managing their savings to young adults navigating their first independent bank accounts.

How modern fraudsters operate

Fake tech support, weapon number one

Fraudsters most often pose as technical support employees, bank representatives or government officials to win their victims' trust, a method that exploits consumers' growing technological confusion with tools they understand less and less, especially as banking apps, verification codes and automated alerts multiply and blur into a constant stream of notifications that few people fully parse before reacting.

Fraudulent websites are involved in 40% of recorded scams, while phone calls, texts and emails play a role in nearly half of reported cases according to the report.

Money sent directly, with no way back

In 49% of scams, victims were led to send money directly through platforms like Zelle or PayPal, instant payment systems that make recovering funds virtually impossible once the transaction goes through.

Artificial intelligence, the new tool of manipulation

Voice cloning and fake endorsements

The report documents the growing use of voice cloning, fake celebrity endorsements and AI-generated impersonations to deceive victims, techniques that make traditional scams infinitely more believable than before, since the audio or video evidence a skeptical victim might once have demanded can now be manufactured convincingly within minutes.

These deepfakes can now simulate the voice of a loved one in distress or that of a company executive giving an urgent order, completely blurring the line between reality and digital manipulation.

A likely undercount of the phenomenon

The report explicitly notes that the real share of AI-driven scams is probably underestimated, since many victims do not even realize artificial intelligence tools were used against them.

The human toll behind the numbers

Significant financial damage for victims

Among recorded victims, 21% report having suffered severe financial hardship as a result of the scam, while an additional 25% report moderate financial hardship, figures that reflect an economic impact far from trivial for affected households, many of whom lost retirement savings or emergency funds they had spent years building up.

Even more striking, 73% of victims say the scam had a negative impact on their mental health or overall well-being, a reminder that these economic crimes leave marks far beyond the bank account.

A trauma too often minimized

This psychological dimension remains largely absent from public debate on fraud, even as the Gallup report confirms that emotional distress affects the vast majority of scam victims, far beyond mere material losses.

Why the West must accelerate its technological response

A race between attackers and defenders

This report confirms a reality that cybersecurity experts have been repeating for years: artificial intelligence benefits criminals as much, if not more, than the companies meant to protect consumers, a worrying asymmetry in the global technology race, since offensive tools tend to spread faster than the defensive systems designed to counter them.

If the West wants to keep its lead in AI development, it must also invest massively in fraud detection tools, or risk seeing this technology turn durably against its own citizens.

The role of payment platforms and banks

Platforms like Zelle and PayPal, heavily used in fraudulent transfers, must strengthen their verification mechanisms, a responsibility these tech companies can no longer afford to keep offloading entirely onto consumers.

What this reveals about consumer vulnerability

A poorly calibrated digital trust

The success of these scams reveals a digital trust that is often poorly calibrated among American consumers, who struggle to distinguish legitimate communication from a sophisticated AI-assisted fraud attempt, particularly when the message arrives at a stressful moment and demands an immediate response.

This vulnerability especially affects people least familiar with new technologies, but the report notes that even savvy users can be fooled by sufficiently elaborate deepfakes.

Digital literacy, the missing link

Faced with this reality, digital literacy education for citizens remains largely insufficient in the United States, a glaring gap even as the tools used by fraudsters become more accessible and more convincing every year.

A national security issue as much as an economic one

Sums that far exceed simple street crime

With $68 billion in annual losses, AI-assisted fraud has reached a scale that far exceeds traditional crime, now resembling a matter of national economic security for the United States, comparable in magnitude to the losses attributed to entire categories of white-collar crime that receive far more sustained law-enforcement attention.

These colossal sums potentially finance organized, sometimes transnational, criminal networks that exploit regulatory gaps between jurisdictions to escape prosecution.

The need for stronger federal coordination

Given the scale of this phenomenon, stronger federal coordination among financial regulatory agencies, cybersecurity authorities and tech companies looks like an urgent necessity rather than merely one option among others.

The regulatory lag against technological acceleration

Laws designed for a pre-AI world

Most American laws governing financial fraud were designed before the explosion of generative artificial intelligence tools, a time lag that leaves regulators constantly one step behind criminals capable of adopting new technologies in a matter of weeks, while legislative reform typically takes years to move through committees and votes.

This regulatory lag is not unique to the United States: most Western democracies struggle to evolve their legal framework as fast as criminals adapt their methods, a chase in which the advantage remains structurally on the side of the fraudsters.

Initiatives that remain too fragmented

Some federal agencies, including the Federal Trade Commission, have started publishing specific alerts about deepfake-related scams, but these initiatives remain scattered and lack the coherent national coordination needed to match the scale of the phenomenon described in this report.

Without a coordinated legislative effort among Congress, financial regulators and tech giants, each new generation of artificial intelligence tools risks widening the gap between the sophistication of fraudsters and the real capacity of institutions to protect America's most vulnerable consumers.

Conclusion: a battle for trust that is only beginning

An alarm bell Washington can no longer ignore

This report from Gallup and the Stop Scams Alliance should serve as an alarm bell strong enough to trigger a serious regulatory mobilization, at a moment when artificial intelligence is transforming fraud methods just as quickly as the legitimate uses of this technology.

Individual vigilance will not be enough

However vigilant American citizens may try to be, the growing sophistication of deepfakes makes it wishful thinking that simple individual caution alone will curb this phenomenon without a structural strengthening of technological and regulatory protections.

The real test of the coming years will be whether American institutions, from federal regulators to the tech companies themselves, can catch up with criminals who, for their part, wait for no legislative approval before deploying their new artificial intelligence tools.

By Maxime Marquette, columnist

Columnist's transparency note

Who I am and my acknowledged biases

I am a columnist, not a cybersecurity expert or a statistician. I believe the West must lead the global technology race, an acknowledged bias that shapes my reading of this report while strictly relying on the data published by Gallup and the Stop Scams Alliance.

What I don't know

I do not know the full methodological detail of this survey beyond what is publicly reported, and I limit myself to the figures confirmed by the sources cited below, without unfounded extrapolation about the real and future scale of a phenomenon still poorly understood even by researchers themselves.

Sources

Primary sources

Yahoo Finance — AI fraud explodes, 15 million Americans scammed in 2025, July 1, 2026

The Hill — AI and deepfakes fuel the scam wave, July 1, 2026

Secondary sources

Fars News — survey finds AI deepfakes driving $68 billion U.S. scam epidemic

Cyber Daily — deepfake scams increase 2,000% in three years

CNBC — technology news

Reuters — technology news