FBI Reports $893M Lost to AI-Powered Scams in 2025

FBI tallies first year of AI scam losses

The FBI's Internet Crime Complaint Center received more than 22,000 complaints tied to artificial intelligence scams in 2025, marking the first year the agency tracked AI-enabled fraud as a separate category. Those complaints represented roughly $893 million in reported losses, according to the bureau's annual report.

Investment scams accounted for $632 million of the total, while Americans aged 60 and older reported $352 million in losses. The figures were first reported by The Conversation, citing analysis by Pawan Jain, a finance professor studying how households make money decisions in the age of AI.

The scams themselves are not new—emergency calls from supposed relatives, phishing emails, fake investment opportunities. What has changed is the cost and quality of execution. Voice cloning now requires only a few seconds of recorded speech, and large language models can generate polished, error-free phishing messages that lack the typos and awkward phrasing that once served as warning signs.

Why it matters

The $893 million figure likely understates the problem. Deloitte projects overall U.S. fraud losses could climb from $12.3 billion in 2023 to $40 billion in 2027, with AI tooling accelerating both the volume and sophistication of attacks. For enterprises, this means legacy fraud detection systems trained on older patterns may miss AI-generated schemes. For families, it means traditional skepticism—"that doesn't sound like my grandson"—no longer works when the voice is an exact match.

How AI changes the fraud economics

Jain argues the core security challenge has shifted from data breaches in server rooms to decision-making at kitchen tables. AI lowers the barrier to entry for scammers by automating tasks that once required skill or scale. A single operator can now run hundreds of personalized phishing campaigns or voice calls simultaneously.

The technology also exploits cognitive vulnerabilities. Fraudsters manufacture urgency—a panicked relative, a time-limited investment, an account that will be locked—because stress narrows attention and pushes people toward fast, intuitive judgments. Payment platforms like Zelle, designed for speed and convenience, make recovery difficult once funds are transferred.

Practical defenses for organizations and households

Jain recommends borrowing risk management practices from financial institutions. The simplest: terminate any urgent call or message and contact the person or organization through a number you already trust, not one provided by the caller.

For families, he suggests creating an emergency code word and treating any urgent request without it as suspicious. Requiring a second person to approve large transfers or imposing a 24-hour delay before sending major payments can disrupt the false urgency scammers depend on.

Jain also points to the United Kingdom's reimbursement rules for scam victims as a potential policy model. When banks must absorb more of the loss, they have stronger incentives to stop fraud before money leaves the account.

The FBI's 2025 data was first reported by The Conversation, with analysis from Pawan Jain.