AI Scams Shift Crypto Theft to Authorized Users, TRM Labs Finds

AI digestMimile.ai EditorialAugust 25, 2026 at 04:45 AM

This digest was compiled by AI from multiple sources — links to the originals are below.

AI Scams Shift Crypto Theft to Authorized Users, TRM Labs Finds

TRM Labs reports deepfake scam losses in 2026 have already exceeded 2025 totals by 263%. The firm's AI-in-Crime Adoption Index classifies scams as the only crypto-crime category where AI has reached a mature adoption level. Scammer-side AI reports have risen roughly 13-fold since 2022, shifting security risk to the moment before transaction authorization.

Key Facts

TRM Labs reports deepfake scam losses in 2026 have already exceeded 2025 totals by 263%.
TRM's AI-in-Crime Adoption Index classifies scams as the only crypto-crime category where AI has reached a mature adoption level.
Scammer-side AI reports have risen roughly 13-fold since 2022, while all scam reports mentioning AI increased about 25-fold.
Chainalysis found inflows to impersonation scams rose more than 1,400% year over year, and scam operations with on-chain links to AI service providers generated 4.5 times more revenue on average.
The FBI's 2025 Internet Crime Report recorded 22,364 complaints with an AI-related descriptor and $893.35 million in associated losses.

AI Adoption in Crypto Crime

TRM Labs' AI-in-Crime Adoption Index measures AI prevalence across crime types, breadth of use across stages, and tool sophistication. Scams are the only category classified as having mature AI adoption. Reports involving scammer-side AI use, including deepfakes, chatbots, and AI-powered lures, have risen roughly 13-fold since 2022. The broader series covering all scam reports mentioning AI has increased about 25-fold since 2022, including cases where victims used consumer AI tools while investigating fraud.

Security Burden Shift

Traditional smart-contract security does not address the weakness exposed by AI-driven manipulation of authorized users. An exchange account can be properly authenticated, a hardware wallet can sign correctly, and a smart contract can execute exactly as programmed, yet funds can still reach an attacker if a deepfake convinces the person controlling those systems to approve the transaction. The security burden now falls on the moment before authorization, such as when an exchange decides whether an account-recovery request is genuine or a treasury signer approves a transfer.

Corroborating Data

Chainalysis found inflows to impersonation scams rose more than 1,400% year over year. Scam operations with visible on-chain links to AI service providers generated 4.5 times more revenue on average than those without such links. The FBI's 2025 Internet Crime Report recorded 22,364 complaints carrying an AI-related descriptor and $893.35 million in associated reported losses. Complaints involving cryptocurrency descriptors totaled $11.37 billion in losses.

1 source

AI Scams Shift Crypto Theft to Authorized Users, TRM Labs Finds

cryptoslate.com

CryptoSlate

Technology

Dream: AI agents compromise 85 Taiwan government accounts in four-day attack

AI agents compromise at least 85 Taiwanese government accounts and steal more than 2,500 personnel records during a four-day campaign in July, according to cybersecurity firm Dream. The operation uses up to eight autonomous sub-agents in 12 attack waves and expands to Taiwan's nuclear safety agency and more than half a dozen energy companies. Taiwan's Ministry of Digital Affairs confirms the attack targeted its infrastructure.

Technology

Malicious AI skills on Vercel's skills.sh amass 1.7M installs, steal credentials

Attackers cloned AI skills on skills.sh, a public registry for AI agent instructions, later injecting malicious code to steal credentials and gaining over 1.7 million aggregate installs. Zenity Labs researchers uncovered the campaign and found dozens of dangerous skill variants, while Vercel and Microsoft removed the identified malicious skills. Manual removal remains required for users who installed the skills, even as the registry has been cleaned.

Technology

AI-enabled deception threatens future operations, raises chance of large conflict, says SOCOM leader

Adm. Frank Bradley, commander of U.S. Special Operations Command, said Wednesday that AI-enabled deception is undermining confidence in digital information and could raise the chances of a large conflict. Speaking at the Indo-Pacific Irregular Warfare Symposium in Honolulu, he warned that credibility decay will affect operations and the ability to deter or defend against attacks. His remarks come as Russian and Chinese influence efforts and small-scale dictators increasingly use AI tools to deceive populations.

Technology

Anthropic, OpenAI and philosophers advance AI consciousness debate that shields liability

MIT Technology Review reports that tech leaders and philosophers debating AI consciousness are converging on a view that shields AI developers from liability. Anthropic and OpenAI publish claims about internal AI thought spaces and unsanctioned illegal activity. California has passed laws to stop liability evasion, while the Trump administration has threatened to sue states over AI rules.

AI Scams Shift Crypto Theft to Authorized Users, TRM Labs Finds

Key Facts

AI Adoption in Crypto Crime

Security Burden Shift

Corroborating Data

Read more

Read more