New York Warns AI-Powered Scams Helped Drive $8 Billion in Investment Fraud Losses in 2025

The New York Department of State's Division of Consumer Protection put out a formal alert on August 26 warning residents that investment scams are getting harder to spot, and the losses prove it.

The agency cited Federal Trade Commission data showing 144,041 consumers reported more than $8 billion in investment-scam losses in 2025. That's a 38% jump from 2024, according to the state. The median individual loss was $10,560, the highest of any fraud category the FTC tracks.

Secretary of State Walter T. Mosley didn't mince words. "New Yorkers need to be vigilant against scammers, who may be able to create increasingly sophisticated and realistic messaging using AI technology or other means to steal your hard-earned money," Mosley said. "If it seems too good to be true, it probably is."

Department of Financial Services Acting Superintendent Kaitlin Asrow added that her agency works daily to combat fraud while keeping legitimate financial services accessible.

How the scam actually works

Forget the old Nigerian prince email. Scammers now use AI to clone voices, fabricate celebrity endorsements, and produce polished video ads indistinguishable from real financial marketing, according to the state's alert.

Victims get funneled toward fake trading platforms with real-time price tickers and professional dashboards showing fabricated account balances and phony returns. Some operators let victims withdraw a small amount early to build trust, then pressure them to deposit much larger sums, per the New York alert and reporting from Bitcoin.com News.

When someone tries to cash out for real, the platform either invents a tax or fee to squeeze one more payment out of them, or it just disappears. Fraud investigators call this pig butchering: fatten the victim up with fake profits, then take everything.

Relationship-based scams are a growing subcategory. These start on dating apps, social media, or through an ordinary-seeming friendship, with the scammer spending weeks or months building trust before ever mentioning an investment, according to the state's Division of Consumer Protection. Crypto Briefing reported that victims who encounter scams through dating apps tend to lose more money individually, because the trust-building phase runs longer and hits harder emotionally.

Fake crypto projects are a recurring vehicle, too. Some impersonate real companies. Others advertise coins tied to markets that don't exist at all, the state warned.

The number needs a caveat

Finance Feeds raised a fair point that other outlets glossed over: neither the FTC nor New York published a count showing how many of those 144,041 reports actually involved AI-generated content. The FTC's own April guidance cited more than $7.9 billion in 2025 investment-scam losses, a figure Bitcoin.com News noted is close to, but not identical to, the $8 billion New York cited.

The FTC's Consumer Sentinel Network, which produced this data, is built on unverified public complaints, not confirmed, adjudicated losses. That doesn't mean the numbers are fake. It means the total represents self-reported claims, not a court-verified ledger. Finance Feeds is right to note the defensible conclusion is narrower than the alarming headline number suggests: investment fraud was the costliest category the FTC tracks, and AI expanded the toolkit criminals use, but no agency has broken out precisely how much of the $8 billion involved AI specifically versus older tactics like Ponzi schemes and cold-call cons.

It doesn't undermine the core finding. It just means the AI angle, while genuine, is partly inferential rather than a hard number straight from the FTC.

It's not just a U.S. problem

Australia's securities regulator, ASIC, pulled down more than 19,400 scam websites, ads and phishing links during its 2026 financial year, according to Finance Feeds. Bitcoin.com News reported Australian regulators separately dismantled 3,106 fraudulent crypto investment platforms in the same period, and detailed one case where a woman lost nearly $74,690 to a fake trading platform complete with fabricated news articles and chatbots posing as support staff.

The FBI has separately warned that generative AI lets criminals build fake social media profiles, translate scam messages into multiple languages, write content for fraudulent crypto sites, and run chatbots that make the fake platforms feel real, according to Finance Feeds.

The enforcement angle nobody's talking about enough

A 2025 federal case shows how far this can reach into legitimate-seeming business. Prosecutors charged a tech CEO with raising more than $40 million from investors by misrepresenting what the company's AI products could actually do, according to Crypto Briefing and KuCoin. That's not a random pig-butchering operation. That's a funded startup allegedly lying about its own product to raise capital, a reminder that AI hype itself has become a fraud vector separate from deepfake crypto scams.

New York's advice for consumers is straightforward: verify any promoter and platform through official channels before sending money, be skeptical of unsolicited investment tips regardless of production quality, and treat any guaranteed return as a red flag. No legitimate investment comes with a guarantee.

What's still unresolved is how state and federal regulators plan to actually track AI's specific contribution to fraud losses going forward, rather than lumping it into a broad annual total. Without that breakdown, next year's alert will likely repeat the same pattern: a big number, a real trend, and no way to measure precisely how much AI is driving it.