Crypto scams just got a high-tech lobotomy, and the tab is absolutely staggering. In 2025, fraudsters armed with AI and deepfake tech pushed estimated losses to a record $17 billion, a figure that would make even Satoshi sweat, according to a new report by Chainalysis.
Turns out, adding artificial intelligence to the theft mix is like strapping a rocket to a dumpster fire. The average scam payment skyrocketed to $2,764 in 2025, up a jaw-dropping 253% from $782 the year prior. Chainalysis Head of Research Eric Jardine notes that over 70% of AI-enabled scams land in the top 50th percentile of transfer volume, meaning they "get bigger faster" than a degen spotting a 100x leverage play.
Scams with on-chain links to AI vendors generated an average of $3.2 million per operation—about 4.5 times more than those without. This pattern is tied to face-swap software and deepfakes sold by Chinese vendors on Telegram. Once scammers can look exactly like someone you know, the believability goes up, allowing them to scale operations in a "really problematic" way that feels like a rug pull on your trust.
Government impersonation exploded by over 1,400% in 2025, with criminals posing as officials from agencies and crypto platforms. One massive campaign targeted U.S. residents with fraudulent "E-ZPass" toll alerts, traced to a Chinese group known as "Darcula." Despite sending 330,000 texts a day, the infrastructure cost less than $500, proving that phishing is apparently the only market with better ROI than Bitcoin.
The "pig-butchering" schemes remain brutal, serving up financial bacon with a side of heartbreak. One San Jose widow lost nearly $1 million after meeting a man named "Ed" on Facebook; she only realized it was a scam after asking ChatGPT if the investment offer made sense. Jardine explains that these long-running relational scams trade off scale for believability, resulting in higher average losses than simple YouTube giveaways, which at least promise a lambo for free.
Impersonation scams are also moving from centralized exchanges to DeFi protocols to launder funds, sliding into DMs like a bad DEX aggregator. Meanwhile, the U.S. Attorney’s Office in Massachusetts is seeking to recover over $200,000 in USDT from a Tinder-based crypto fraud, proving that swiping right on a bad trade can cost you more than just gas fees.
While basic automation moves funds on-chain, advanced AI is increasingly used to create fake, KYC-compliant exchange accounts to cash out. This sustains the physical "scam compounds" in Myanmar and Cambodia, where human trafficking and forced labor fuel the industry. Chainalysis concludes that these operations reveal a dark human cost, turning "pig butchering" into a massive industry run by organized crime, where the only thing getting slaughtered is human dignity.
