National Tax Service Deploys AI to Hunt Crypto Tax Evasion Using Virtual Asset Analysis

(SOUTH KOREA) — South Korea’s National Tax Service is developing a comprehensive virtual asset integrated analysis system that uses artificial intelligence to track cryptocurrency transactions and detect tax evasion.

The National Tax Service framed the rollout as a move from pilot testing throughout 2024 to full-scale operations beginning in 2025, signaling a shift toward analytics-led enforcement as crypto transactions grow harder to trace across platforms.

By building the system around direct exchange data and blockchain tracing, the tax authority is aiming to shorten the distance between raw trading records and audit-ready leads. The approach also raises compliance expectations for registered exchanges that already provide data to regulators.

South Korea’s plan centers on combining multiple data sources into a single analytical environment rather than treating exchange records and on-chain activity as separate views of a taxpayer’s behavior. The system establishes direct data feeds from registered cryptocurrency exchanges and incorporates blockchain analysis tools for transaction tracing.

Those components feed AI-driven detection intended to surface patterns that may indicate tax evasion, while linking the results to existing tax administration databases for what the tax authority described as comprehensive taxpayer profiling. Instead of focusing on a single venue, the platform is designed to analyze transaction patterns across multiple platforms simultaneously.

That cross-platform design targets strategies in which funds move quickly between exchanges, sometimes in arbitrage-style paths that can obscure the original source of gains. The National Tax Service also expects the system to identify potential offshore transfers, where detection can depend on connecting activity that looks separate in any one dataset.

Mechanically, the system’s backbone is data ingestion from more than one pipeline at once, starting with direct exchange feeds and extending to on-chain records that blockchain tools can trace through transaction flows. The tracing element is meant to help investigators follow wallet linkages and see how assets move between addresses and services.

On top of that data foundation, the National Tax Service plans to apply artificial intelligence algorithms that flag suspicious patterns and produce risk profiles for follow-up work. The models are expected to focus on behavior that can look like layering, rapid cycling, or cross-platform patterns that repeat across accounts and venues.

The platform is also designed to process real-time transaction data from multiple sources, which puts pressure on both speed and control. The tax authority described requirements to maintain data security while keeping the information accessible to authorized tax officials, pointing to access controls and oversight expectations that would accompany any broader use inside the agency.

Alongside the technology build-out, the National Tax Service announced plans to create a specialized unit dedicated to managing crypto tax evasion activities. The division appears in the NTS’s 2026 National Tax Administration Operation Plan.

The planned unit is expected to manage the new system and turn its analytics into case development, giving the agency a way to run sustained oversight rather than relying on one-off audits. Its responsibilities also include coordinating with broader tax enforcement functions as the system feeds leads into existing investigative workflows.

South Korea’s move comes as other regulators expand their own use of artificial intelligence to monitor crypto markets, increasing the chances that suspicious activity triggers scrutiny from more than one direction. The Financial Supervisory Service released its 2026 crypto oversight plan on February 9, setting out AI-based monitoring systems to detect market manipulation and suspicious trading behavior.

Under the Financial Supervisory Service plan, AI tools scan trading activity and text signals to detect unusual price jumps within minutes and identify linked wallet groups acting together. Combined with the National Tax Service’s transaction-tracing and risk-flagging approach, the parallel build-outs increase pressure on exchanges and high-volume traders to ensure internal controls match the level of monitoring authorities say they are preparing to deploy.