Economic Essay Contest AI data sovereignty: The case for a local server interceptor in Korean finance The Screenshot Nobody Thinks About An elderly woman in Seoul takes a photo of her bank statement. She is confusing about a transaction. Her grandson told her to “just ask ChatGPT.” She uploads the image. Within seconds, the AI explains the charge. Problem solved. Except it is not solved. That bank statement with her account number, transaction history, and personal details just traveled to a server in California. She has no idea. Her bank has no idea. And until something goes wrong, nobody will care. This is happening right now, thousands of times a day, across Korea’s financial system. We have no infrastructure to stop it. The Invisible Data Leak Financial institutions are racing to adopt AI tools. Chatbots answer customer questions. AI assistants help employees draft reports. These tools are useful. Banning them would be impossible. But there is a silent risk: When users interact with AI systems, data flows outward. A customer uploads a document. An employee pastes a client email into a writing assistant. A researcher screenshots proprietary data. Each action sends information to foreign servers. Most AI tools — ChatGPT, Google Bard, Microsoft Copilot — process data in data centers outside Korea. Even with encryption, Korean financial data is leaving Korean jurisdiction. The current regulatory approach is to tell people “do not upload sensitive information.” That is not a policy. It is a hope. Not Just the Elderly, Everyone Is at Risk It would be easy to frame this as a problem only for elderly or less tech-savvy users. But the data tells a different story. Studies show that people across all age groups and education levels routinely share sensitive information with AI tools without understanding where that data goes.