Interior ministry finalizes AI ethics standards for public sector

The interior ministry said Thursday it has finalized an ethics standard for artificial intelligence in the public sector, setting out six core values and 68 self-check items that government bodies will apply according to how deeply they use AI.

The standard, drawn up under the national task of building what the government calls the world's best AI-powered democracy, is meant to address concerns such as biased decisions and privacy infringement as AI becomes a working partner for public officials.

Six values anchor the document: publicness, equity, transparency, accountability, safety and privacy protection. The 68 checklist items translate those values into practice, applied through a three-stage self-check system matched to the level at which each service relies on AI, from basic use through decision support to autonomous decision-making.

For work that affects citizens' rights — selecting welfare beneficiaries, granting permits and licenses, or predicting risk — the ministry said final responsibility and control must rest with people, not machines, and agencies must establish human oversight, follow-up management and redress procedures.

The ministry described the standard as a practical tool rather than a declaration of values. It plans training programs for public-sector employees and a feedback system to catch new ethical issues early.

"As AI technology becomes a core driver of administrative innovation, an ethical foundation to protect citizens' basic rights and secure trust is absolutely necessary beneath it," Interior Minister Yoon Ho-jung said. "Through the ethics standard and the 68 self-check items, we will make AI ethics settle naturally across the public sector and build an administrative environment where AI can be used fairly and safely."