FBI’s 2025 Internet Crime Report reveals cryptocurrency and AI scams bilked Americans of billions this year

The FBI’s Internet Crime Complaint Center released its 2025 annual report a few weeks ago, and the numbers reveal a criminal landscape transformed by artificial intelligence and cryptocurrency—two technologies that have become the preferred tools for bilking Americans at scale.

The central finding cuts across traditional cybercrime categories: cryptocurrency and AI-powered scams have become the dominant cost drivers in online fraud. This isn’t a marginal shift. These attack vectors now represent billions of dollars in losses, a jump that signals how quickly criminal infrastructure has adapted to exploit emerging technologies. For the average American, it means the threat model has changed—the scammer on the other end of a phishing email is now likely using AI to craft more convincing messages, and the money they steal is increasingly funneled through cryptocurrency networks designed to obscure its trail.

The 2025 report, published by the FBI’s Internet Crime Complaint Center (IC3), aggregates complaints and loss data from across the United States. While the source materials do not itemize every category of loss or provide granular breakdowns for this article, the headline finding is unambiguous: cryptocurrency and AI scams have become the primary financial hemorrhage in internet crime. The scale of these losses—measured in billions—represents a seismic shift from the ransomware-dominated threat landscape of previous years.

What makes this moment significant is the convergence of two technologies that criminals have weaponized in tandem. AI tools can now generate convincing deepfake videos, craft personalized phishing emails that mimic trusted contacts, and automate social engineering at scale. Cryptocurrency provides the exit ramp—a way to move stolen funds across borders and through decentralized networks without traditional banking oversight. Together, they form a nearly frictionless crime pipeline.

How Do Modern AI Scams Mirror Cambridge Analytica’s Playbook?

The parallel to Cambridge Analytica is instructive here, though the mechanics differ. Cambridge Analytica harvested personal data at scale—psychological profiles, behavioral patterns, consumer preferences—and used that data to micro-target voters with personalized messaging designed to manipulate behavior. The firm didn’t need to steal money; it stole attention and agency. Today’s AI-powered scammers are doing something structurally similar: they’re harvesting behavioral data (how you respond to urgency, social proof, authority figures), using AI to personalize their attack vector, and exploiting psychological vulnerabilities at scale. The difference is the endpoint: Cambridge Analytica aimed at political outcomes; today’s scammers aim at your bank account. But the underlying mechanism—behavioral inference, micro-targeting, consent erosion—remains the same. Both operate in the gap between what users think they’re sharing and what’s actually being extracted.

The FBI’s report also serves as a public warning about the speed at which criminal tactics evolve. Ransomware dominated headlines for years. Now it’s being displaced by AI-generated fraud and cryptocurrency theft. This suggests that defenders are always fighting the last war while criminals are preparing for the next one. The criminals with access to AI tools can now operate with less human intervention, lower overhead, and higher success rates. A single compromised email account can become the launching point for dozens of AI-generated phishing campaigns, each one tailored to the recipient’s known interests and vulnerabilities.

Why Are Traditional Defenses Failing Against AI-Enhanced Attacks?

For individuals, the practical implication is clear: the traditional defenses—skepticism toward unsolicited emails, verification of sender addresses, strong passwords—are becoming insufficient. An AI-generated deepfake video of your boss requesting an urgent wire transfer is far more convincing than a typo-laden phishing email. A personalized message referencing details from your LinkedIn profile or recent purchase history carries more social engineering weight than a generic “confirm your password” request.

The lessons from Cambridge Analytica become relevant here: when personal data is weaponized for behavioral manipulation, the attack succeeds not through technical sophistication but through psychological precision. Christopher Wylie, the whistleblower who exposed Cambridge Analytica’s methods, revealed how psychological profiling could predict and influence individual behavior. Today’s AI scammers are applying similar principles—they’re not just sending random phishing emails, they’re crafting messages designed to exploit specific psychological triggers based on harvested behavioral data.

What Does the Cryptocurrency Connection Mean for Recovery?

The FBI’s 2025 Internet Crime Report does not specify what preventive measures individuals should take, but the underlying data suggests that the threat model has fundamentally changed. Cryptocurrency scams in particular exploit the irreversibility of blockchain transactions—once funds are sent, they cannot be recalled. AI scams exploit the erosion of trust in digital communication itself. When you cannot reliably distinguish a video of someone you know from a deepfake, or a personalized message from a human contact versus an AI-generated impersonation, the friction that once protected you has vanished.

The full 2025 Internet Crime Report is available through the IC3’s official channels. As these crime categories continue to evolve, the question facing regulators and technologists is whether AI and cryptocurrency guardrails can be built faster than criminals can exploit them.