AI Turned Crypto Fraud Into a $30 Billion Assembly Line

The Numbers

TRM Labs tracks illicit cryptocurrency flows across the globe. Their 2025 report puts total illicit crypto volume at a record $158 billion, a 145% year-over-year increase. Scam-related activity alone accounted for an estimated $30 billion of that, and generative AI is the main reason the numbers jumped so fast.

What Changed Between 2022 and 2025

In 2022, crypto scams were mostly human-operated. A person behind a keyboard, crafting a message, pretending to be someone else. It was slow and had real limits.

By 2025, that model was obsolete. The share of crypto scam reports involving AI elements like deepfakes and automated chatbots grew approximately 13 times since 2022, according to TRM Labs. Scammers now deploy tools that generate synthetic identities, automate conversations, clone voices, and route stolen funds with minimal human involvement.

How the Scams Actually Work

The AI-powered fraud patterns TRM Labs describes fall into a few recurring categories.

Fake identity creation. Generative AI produces realistic profile photos, backstories, documents, and video. Fraudsters use these synthetic identities to build trust before asking for money. A "financial advisor" with a LinkedIn profile, a headshot, and three years of fake posts feels real. That's the point.

Voice and video cloning. Voice cloning now requires as little as 20 to 30 seconds of audio. Once a scammer has a sample, they can generate a convincing fake call in the voice of someone you know. In February 2024, engineering firm Arup lost $25 million when an employee joined a video call where every participant was a deepfake. The faces looked right. The voices sounded right. Nothing was real.

Automated chatbots running at scale. Where a human scammer could manage a handful of targets at once, an AI chatbot can run thousands of conversations simultaneously, responding in seconds and adjusting its pitch based on replies. This is why the volume numbers are so large. It's not more scammers doing more work; it's the same number of scammers doing exponentially more damage.

Laundering automation. Once money moves, AI tools route it through complex chains of transactions designed to obscure its origin, with minimal oversight, making recovery nearly impossible.

Who Gets Targeted

Everyone, but some groups more than others. People unfamiliar with crypto mechanics are frequent targets because they don't know what a legitimate transaction looks like. Older adults get hit through romance scams and fake investment platforms. Younger people get hit through social media impersonation and fake trading opportunities.

Every scam works by convincing you that you know and trust the person or institution reaching out. AI makes that imitation faster and more convincing than it's ever been.

The Impersonation Problem

The $30 billion in scam losses isn't really a crypto problem. It's an identity problem. People lost money because they believed someone was who they claimed to be: a trusted friend, a family member in trouble, a financial institution, an employer.

A cloned voice sounds like your son. A deepfake video looks like your financial advisor. An automated chatbot types like a customer service rep. Human detection accuracy for AI-generated voice and video hovers around 55 to 60%, barely better than a coin flip. We can't reliably spot these fakes on our own.

What Stops a Fake Voice

AI voice cloning can copy how someone sounds. It can't know a secret.

Trust Onion gives families a set of three rotating codewords that change every 60 seconds, calculated locally on each person's phone with no server required. When someone calls claiming to be a family member, you ask one question: "What are the words?"

A cloned voice doesn't know the words. A scammer running an automated call doesn't know the words. Nobody outside your family knows them, because they're never transmitted anywhere. The words rotate on a schedule, so even if someone overheard them once, they'd expire within a minute.

It's free to use, works offline, and takes about three seconds to stop a scam call that might otherwise cost someone their savings.

What You Can Do Today

Talk to your family about how these scams work: the fake urgency, the request to keep it secret, the pressure to move fast. These patterns repeat across scam types.

Set up a verification system before you need one. Don't wait until you get a call from someone who sounds exactly like your daughter saying she's in trouble and needs money right now. By then, the pressure is already on.

Be careful with any request that involves cryptocurrency. Legitimate institutions don't demand payment in crypto, and family members who are genuinely in trouble have other options.

TRM Labs data shows this problem is getting worse. The tools scammers use are getting better and cheaper, and the gap between what AI can fake and what humans can detect keeps growing. Your defense doesn't have to be technical. It just has to be shared.