- TRM says AI adoption across crypto crime rose 40% within one year.
- Deepfake scam losses in 2026 exceeded the full-year 2025 total by 263%.
- Coldcard and Bybit are using AI to find flaws and stop suspicious transfers.
AI crypto crime adoption rose 40% over the past year, according to TRM Labs. Its 2026 index scored use at 54 out of 100, up from roughly 28 in 2024. Scams reached a mature adoption stage, while hacking and ransomware remained emerging. Security teams are also using AI to expose flaws and block suspicious transfers.
AI Crypto Crime Expands Through Scams and Automated Attacks
TRM said AI-linked crypto scam reports increased up to thirteenfold since 2022. Reported deepfake scam losses during 2026 already exceeded the full-year 2025 total by 263%. Chatbots, synthetic identities, and cloned media let smaller groups target more victims.
AI crypto crime also entered automated ransomware operations. Sysdig documented JADEPUFFER, an AI-driven campaign targeting an exposed Langflow server. The agent performed reconnaissance, credential searches, lateral movement, and destructive database encryption. It corrected one failed login within 31 seconds, showing how quickly an automated attack can adapt.
TRM counted 201 digital-asset hacks during the first half of 2026. About 75% of losses came from only 4% of incidents, mostly infrastructure compromises. North Korea-linked operations accounted for roughly $600 million, or 61% of losses.
AI Crypto Crime Pushes Wallets and Exchanges Toward Automation
In a recent blog post, Coinkite used Kimi and other frontier models to review Coldcard firmware after a seed-generation flaw. Version 5.6.1 requires users to add physical randomness and adds transaction-integrity checks.Â
New seeds require 65 unpredictable key presses, 50 die rolls, or 128 coin flips. Customers with affected seeds must create new ones and move their funds. Installing the update alone cannot secure an already compromised wallet.
Bybit said AI-supported controls stopped more than 30,000 suspicious withdrawals during the first half. The exchange estimated those requests represented over $700 million in potential user losses. That amount describes prevented potential losses, not stolen funds later recovered. Its automated audits found serious flaws at three to five times the manual review rate. AI tools also helped process more than 100,000 security alerts.
Against AI crypto crime, Bybit said automated testing cut assessment cycles from two weeks to two hours. The exchange also blacklisted more than 10,000 addresses linked to suspected malicious activity.
Disclaimer: This article is for informational purposes only and does not constitute financial advice. CoinCryptoNewz is not responsible for any losses incurred. Readers should do their own research before making financial decisions.
