Artificial intelligence is spreading quickly across crypto-related crime, and TRM Labs says scam networks are leading the change. The blockchain intelligence firm reported that AI adoption in crypto crime rose 40% over the past year.
In TRM Labs’ 2026 AI Adoption Index for Criminal Activity, overall AI usage in crypto crime climbed from a score of 28 in 2024 to 54 out of 100, placing the sector in the firm’s “Emerging” category.
Scams rank highest in AI adoption
By segment, scam activity has already reached the “Mature” stage in AI adoption. Hacking operations and ransomware remain in the “Emerging” stage, while drug trafficking and darknet markets are still in the “Horizon” stage.
Ari Redbord, TRM Labs’ global head of policy and government affairs, said: 「AI isn’t creating new forms of crime. It is removing the constraints that once limited traditional crime.」
According to Redbord, AI has lowered the technical barrier for criminal activity while expanding the scale of attacks. Identity fraud has also moved toward industrialized mass production. Tasks that once required a full hacking team can now be handled by a single person with a paid AI subscription, he said.
Reports tied to deepfakes and AI chatbots rose 13x since 2022
TRM Labs also said reports of crypto scams involving deepfake technology or AI chatbots have increased 13-fold since 2022.
The firm added that financial losses from deepfake scams so far in 2026 are already 263% higher than the full-year total in 2025. In its view, AI is rapidly becoming a core tool for criminal groups looking to scale fraud operations.
North Korea-linked groups highlighted in the report
TRM Labs said North Korean hacking groups are making extensive use of deepfakes to pose as IT workers and infiltrate companies. The report said they are also using AI for social engineering attacks and to automatically identify code vulnerabilities that can be used against crypto companies and blockchain protocols.
Several security experts have recently warned that AI can identify vulnerabilities that humans might miss, including flaws that are difficult to detect.
Speaking at the Wyoming Blockchain Symposium in 2026, Global Settlement Network CEO Ryan Kirkley said AI agents could soon crack Wi‑Fi networks, passwords and crypto wallets at a scale not seen before.
He said: 「In the past, it wasn’t worth the effort to hack a retail user holding just $20,000 in assets. Now I can deploy an army of AI agents and attack everyone at once.」
Zcash flaw and record hack count in H1 2026
The concern has already moved beyond theory. In June this year, security engineer Taylor Hornby used AI to discover a critical vulnerability in the Orchard transaction pool of privacy coin project Zcash. If exploited, the flaw could have allowed an attacker to mint an unlimited number of counterfeit tokens inside the pool.
TRM Labs said digital asset hacks hit a record 201 incidents in the first half of 2026, more than double the level seen in 2025. It added that 75% of losses were concentrated in just 4% of major cases, and most of those incidents involved infrastructure weaknesses tied to stolen private keys or compromised credentials.
Activity linked to North Korea alone accounted for about $600 million in stolen funds, or roughly 61% of total losses in the first half, according to the report.