South Korean fraudsters jailed for AI cryptocurrency scam
Between 2017 and 2019, fraudsters in South Korea deceived investors by claiming an AI system could profit from cryptocurrency arbitrage across global exchanges. They collected approximately 19 billion won from roughly 2,800 investment transactions before being convicted; an appellate court upheld imprisonment sentences in January 2026.
What happened
Between 2017 and 2019, a group led by individual A and including recruiters B and three others received approximately 19 billion won from investors across roughly 2,800 transactions under the pretext of cryptocurrency investment. They claimed an AI-equipped computer connected 7,000 global cryptocurrency exchanges to execute profitable arbitrage trades by buying Bitcoin at low prices in one country and selling at higher prices in another. However, the computer system they presented lacked this capability, and the 'dollar' currency investors received was merely a point with virtually no monetary value.
Reported impact
Investors were deceived by promises of guaranteed high returns and invested without properly investigating the actual mechanism. The court noted that even after the scheme's illegality was exposed, the perpetrators continued actively recruiting investors, expanding the harm.
Index analysis
The fraud centered on misrepresenting AI capabilities and cryptocurrency arbitrage operations to solicit investments. Investors received worthless points instead of actual trading gains, and the promised AI system did not exist or function as described.
Governance controls
The Jeonju District Court's appellate division sentenced A to 3 years and 6 months imprisonment and co-conspirators B and three others to sentences ranging from 10 months with 2 years suspended to 1-2 years, taking into account the scale of proceeds, continued recruitment after discovery, and investor vulnerability to high-return claims.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.