AI Adoption in Crypto Crime Jumps 40%, TRM Labs Index Shows

Scammers account for most of the growth, TRM Labs said. The proportion of cryptocurrency scam reports involving AI elements such as deepfakes or automated chatbots has increased by as much as thirteen times since 2022. AI allows scammers to generate personalized content at scale, TRM Labs said, using deepfakes to impersonate trusted figures and chatbots to interact with victims. Losses attributed to deepfake-related scams in 2026 so far have already exceeded the entire 2025 total by 263 percent.

Hackers are incorporating AI more aggressively, TRM Labs said. North Korean-linked cyber actors are using deepfake-assisted social engineering to infiltrate organizations through fake IT worker profiles, and using AI to identify software vulnerabilities. AI helps automate reconnaissance and vulnerability scanning, TRM Labs said, making it easier for attackers to find entry points. Digital asset hacks hit a record 201 incidents in the first half of 2026, more than double the figure from the comparable period the year before, TRM Labs said. Roughly 61 percent of the associated losses, about $600 million, were tied to North Korea-linked activity, the firm said. A large share of total losses stemmed from a small minority of incidents that typically involved compromises of private keys or credentials rather than purely code-based exploits, TRM Labs said.

TRM Labs also said it identified a development it called JadePuffer, described as the first fully agentic ransomware attack. In that case, an AI system managed the entire process from reconnaissance and credential theft through lateral movement and encryption without continuous human direction. The firm said the JadePuffer case shows a shift toward fully autonomous ransomware attacks, where AI handles the entire lifecycle without human intervention. TRM Labs also said that ready-made, no-code ransomware packages are now available for relatively low prices, further lowering entry barriers.

Because a significant portion of illicit proceeds eventually moves across public blockchains, on-chain patterns serve as a useful indicator of broader trends in AI-enabled crime, the firm said.

Deepfake scam losses in 2026 have already exceeded the 2025 total by 263 percent, TRM Labs said.