Jeong Yeon-tae AI Lab Designs ‘OXMV GPT’ for Integrated Tax Case Verification

The Jeong Yeon-tae AI Research Institute has developed an artificial intelligence system called the "Korea Tax and Taxation OXMV GPT" for analyzing complex tax cases. This system goes beyond categorizing tax issues and instead examines the underlying facts, such as transaction parties, fund flows, assets, timing, applicable laws, evidence, and procedural requirements.

The OXMV GPT framework includes ontology, explainable AI, and multiple verification stages. The ontology stage structures the transaction details, including who performed what action with whom, when, and the relationship between the parties. The explainable AI stage determines why specific taxes and legal provisions may apply. The multiple-verification stage cross-checks statutes, enforcement decrees, regulations, official interpretations, Tax Tribunal decisions, and court precedents to ensure accuracy.

The system pays special attention to timing, as tax laws can change over time. It distinguishes between current laws and the ones in effect when an event occurred, and examines any supplementary provisions or transitional measures that may be relevant. Additionally, it identifies the law's effect on personal use of corporate funds, employee or freelancer income, and potential civil or criminal issues that may arise.

The Evidence Ledger records the connections between claims and supporting material, ensuring a traceable path from the conclusion to the evidence used. This allows users to understand how a tax conclusion was reached, which can be particularly helpful in tax litigation. The system separates a judgment into various components, such as case number, court, decision date, ruling, recognized facts, arguments, applicable law, judicial reasoning, decisive factors, and appeal status. This helps determine if a precedent is factually and legally comparable to the user's case.