TRM Labs Reports a 40% Year-on-Year Surge in Criminal AI Adoption for 2026

According to data from blockchain intelligence firm TRM Labs, criminal adoption of artificial intelligence (AI) climbed 40% year-on-year in 2026. TRM's new AI-in-Crime Adoption Index places overall adoption at a score of 54 out of 100, marking a significant increase from a score of about 28 in 2024.

Scams Lead the Way in Maturity

Scams are currently the only category rated as "Mature" by the index, driving the majority of growth in criminal AI usage. TRM reports that AI now manages the entire lifecycle of a scam, generating target lists and lures, powering deepfakes, and executing live conversations with victims.

The share of scam reports involving AI has increased roughly 13 times since 2022. Furthermore, 17% of crypto scams with live domains currently advertise AI products, while reported deepfake scam losses in 2026 have already surpassed the total for 2025 by 263%. Conversely, narcotics sit at the opposite end of the spectrum, with darknet buyers actively warning each other away from suspected AI-generated markets.

Hacks and Ransomware Advance Across the Attack Chain

Hacking and state-sponsored theft rank one tier lower at the "Emerging" level. TRM recorded 201 hacks during the first half of 2026, compared to 83 incidents a year prior. Notably, 4% of these incidents accounted for 75% of the total stolen value, with North Korea responsible for approximately $600 million—or 61%—of the half-year total. Two major April operations drove this figure: the $285 million Drift Protocol breach and the $292 million KelpDAO exploit, both of which began with social engineering rather than novel code.

Ransomware has advanced even further along the attack chain. No-code ransomware kits are now traded for $400 to $1,200, and July saw the documentation of JadePuffer, described as the first fully agentic ransomware. This AI agent managed reconnaissance, credential theft, lateral movement, and encryption independently without human direction.

Outlook for Defense and Enforcement

TRM's findings highlight that AI touches every stage of the crime lifecycle, lowering entry barriers while increasing attack scale and sophistication. Maintaining parity between offense and defense will depend heavily on whether enforcement and compliance tooling can scale at the same pace.