FATF warns crypto laws are advancing globally, but enforcement is lagging as AI-driven crime and freeze-resistant stablecoins rise.
The Financial Action Task Force has released its seventh compliance report card on virtual assets. Published July 16, the update tracks how well nations have adopted the FATF’s R.15 recommendations since they launched over seven years ago.
According to a report by Chainalysis, global crypto laws are advancing steadily, but enforcement has not kept pace. Criminal networks are now leaning on artificial intelligence and issuing their own “freeze-resistant” stablecoins to dodge existing controls.
The message from the watchdog is blunt: paper compliance no longer cuts it.
Legislation Grows While Enforcement Stalls
Progress on paper looks strong. Chainalysis reports that 86% of the 147 surveyed jurisdictions have completed virtual asset risk assessments, up from 76% in 2025.
Travel Rule legislation now covers 83% of jurisdictions, compared to 73% a year earlier. Jurisdictions rated “Largely Compliant” climbed from 29% to 34%.
Among the 95 jurisdictions requiring VASP licensing, 81% now run supervisory inspections and 71% have taken enforcement action.
Yet cracks remain visible beneath those numbers. Sixty percent of jurisdictions with Travel Rule laws have taken no supervisory or enforcement action on them.
When FATF assessed preventive AML and CFT measures, only 13 of 139 jurisdictions, fewer than 10%, fully met the standard.
Licensing tells a similar story. While 73% of jurisdictions require VASP licensing, only 58% have actually issued one, and just 40% satisfactorily meet the criterion in mutual evaluations.
Prohibition without follow-through is also spreading, with 23% of jurisdictions now banning VASPs, up from 11% in 2023, without corresponding enforcement gains.
The FATF just dropped its 7th Crypto Compliance Report Card. The main takeaway is that global crypto laws are advancing, but enforcement is lagging.
With criminals leveraging AI and issuing “freeze-resistant” stablecoins to evade controls, paper-only compliance is no longer…
— Chainalysis (@chainalysis) July 23, 2026
Freeze-Resistant Stablecoins Signal a New Threat
One case detailed in the report stands out.
A Cambodia-based conglomerate laundered at least $4 billion between August 2021 and January 2025, according to Chainalysis, linking organized fraud, underground banking, and blockchain-based laundering. At least $37 million of those funds traced back to DPRK cyber heists tied to weapons programs.
After a stablecoin issuer froze over $29 million belonging to that network, the group launched its own USD-pegged token marketed as immune to asset freezing, issued across several public blockchains plus a proprietary chain.
FATF warns that VASPs may no longer be able to rely on issuer-level freeze or burn functions as a safeguard. Terrorist groups including ISIL and Al-Qaeda are also shifting toward stablecoins over Bitcoin for fundraising and transfers, the report notes.
Chainalysis data shows stablecoins now account for 84% of all illicit transaction volume tracked across the industry.
AI and Convergence Push Regulators to Adapt
FATF describes artificial intelligence as a structural factor amplifying money laundering, terrorist financing, and sanctions evasion risks, rather than a standalone threat.
Cited cases include deepfake recruitment scams that stole over $1 million, AI-assisted development of smart-contract exploits, and open-weight models used to bypass commercial safety guardrails. Chainalysis separately found that AI impersonation scams were the fastest-growing fraud category over the past year.
The report also flags DeFi as a persistent blind spot, with 93% of jurisdictions yet to identify qualifying DeFi arrangements that could fall under VASP rules.
Offshore VASPs remain another gap, actively soliciting customers and advising VPN use to mask their operations.
FATF’s recommendations now list blockchain analytics, wallet screening, and blacklisting tools as baseline expectations for any compliant AML and CFT framework going forward.
Be the first to comment