Korean Police Agency, ThinkforBL Complete AI Risk Management Framework

The Korean National Police Agency's Future Policing Policy Bureau has completed a joint research project with AI reliability company ThinkforBL to establish standards for AI impact assessment and risk management, according to datanet.co.kr.

ThinkforBL reviewed the risk levels of 29 AI systems that the police agency operates or plans to introduce. The project also set up an integrated organization-level system to manage AI risks continuously across planning, development and operation.

The framework covers determining AI adoption suitability and risks at the ordering stage, verifying reliability criteria during development, and checking appropriateness during operation. It connects AI impact assessment, risk management policy, organizational structure, procurement guidelines, and safety and reliability documentation.

Under the system, AI operations can be suspended when serious problems occur, and records of the basis for risk judgments are required even when managers change. The framework reflects the EU AI Act, ISO international standards and the NIST AI Risk Management Framework alongside domestic AI law.

ThinkforBL presented the results at the 2026 Policing AI Joint Academic Seminar. The final research report is due for release in November.

Background

The police agency and ThinkforBL began the joint research in July. The AI Data Administration Act, which requires public institutions to conduct AI impact assessments and risk management measures, takes effect in February next year.