The alarming rise of AI-powered fraud
I was looking at the TRM Labs report from this week, and the numbers are pretty staggering. They found that large language model use in crypto scams jumped five times in 2025. That’s not just a small increase—it’s a massive shift in how fraudsters operate.
What’s happening is that scammers are using AI to make their schemes more convincing. They’re creating fake personas with generated images, cloning voices, making deepfake videos. All of this makes it cheaper and easier to trick people. The report mentions how these tools help scams cross language and cultural barriers with less friction.
There was that case last March where at least three crypto founders reported stopping North Korean hackers who used deepfakes in fake Zoom calls. That’s the kind of sophistication we’re dealing with now.
Scammers are combining tactics
Another worrying trend is how different scam types are converging. It’s not just one approach anymore. TRM Labs describes how scammers might start with a romance scam to build trust, then switch to offering fake investments, and finish with a tax scam demanding payment for non-existent fees.
Victim journeys are getting longer and more complex. People might think they’re just talking to someone online, but they’re actually being led through multiple phases of deception. The report calls this a defining trend in crypto fraud today.
Industrial-scale operations
What’s really concerning is how businesslike these operations have become. Scam networks are using efficient structures, specialization, and standardized playbooks. They’re targeting victims at scale, almost like legitimate businesses would approach marketing campaigns.
Behind the scenes, there’s a whole ecosystem of illicit service providers. Some offer AI-as-a-service tools to automate outreach. Others sell phishing kits or provide access to breached data. These services lower the barrier to entry for fraud actors and let them replicate scams across different regions.
The bigger picture of crypto crime
TRM Labs estimates that illicit crypto wallets received about $158 billion in value last year. That’s a 146% increase from 2024’s $64 billion. Part of this jump comes from better monitoring technology revealing more activity, and increased sanctions against countries like Russia.
Interestingly, while the illicit volume went up, its proportion of overall crypto volume actually fell slightly—from 1.3% in 2024 to 1.2% in 2025. So legitimate crypto activity is growing even faster than the criminal side, but the absolute numbers are still huge.
About $35 billion in cryptocurrency went to scammer addresses in 2025. That’s actually a slight decrease from $38 billion the previous year, which might suggest some progress in detection and prevention, but it’s still an enormous amount.
The report makes it clear that we’re dealing with a new era of crypto crime. It’s not just individual hackers anymore—it’s organized, industrialized operations using the latest technology. And with AI tools becoming more accessible, this problem might only get worse before it gets better.
I think what we need is better education for users, stronger security measures from platforms, and continued development of detection tools. But it’s going to be an ongoing battle as scammers keep adapting their methods.
