How AI is Transforming South Korea’s Police Force: Key Innovations and Future Plans
Daniel Kim Views
|
What began as a seven-person task force has expanded into an agency-wide AI effort at South Korea’s National Police Agency. As the governance arm, the new unit plans to produce a comprehensive implementation plan, identify priority projects and set ethical standards — positioning itself as the agency’s official AI control tower.
Lee Chi-hwa, a superintendent who has spent 18 years handling planning, organizational work and AI initiatives within the agency’s Future Public Safety Policy Bureau, told Asia Today on April 17 that the division is prepared to lead those efforts.
The National Police Agency recently reorganized the Future Public Safety Policy Bureau to sharpen its AI focus, aligning with the government’s broader push to accelerate AI development. The department that houses the AI Policy Division was renamed and retooled to prioritize AI, and the AI Policy Division has been assigned the central role of overseeing AI policy, governance and related legislation. Lee said the reorganization reflects how the work has scaled: seven staffers launched the Public Safety AI Innovation task force in May last year; those personnel carried the initial workload and, with the reorganization, the agency is now addressing AI at the institutional level.
Lee outlined the unit’s top priorities: drafting a comprehensive plan to implement AI initiatives designated as national objectives, issuing internal directives that embed ethical guidelines, and delivering in-house training to raise AI literacy across the agency.
He said the agency finalized a Public Safety AI Innovation comprehensive plan in February and designed it so every function within the National Police Agency participates. The plan has been in effect for about three months; the unit completed its first review meeting and will assess progress quarterly. Lee added that the agency designated 12 flagship projects to lead implementation — including an AI-driven public-service initiative called “Everyone’s Police Officer,” targeted to begin operating next year — and has selected 25 projects from a broader pool for active review and advancement.
On ethics and oversight, Lee said the agency will create internal ethical guidelines and codify them in formal directives to guard against abuses such as unauthorized surveillance. The AI Policy Division will also receive explicit authority to coordinate departmental roles on AI work, enabling it to steer agency-wide AI policy. Lee said the directive should be issued in the first half of this year, after which the governance unit can function fully as the agency’s control tower and accelerate AI policies and projects.
Lee identified automated de-identification of sensitive data as a priority technical gap. Police datasets contain a high volume of sensitive information, and if de-identification is complex or costly, it hinders AI development and deployment. Automating that process, he argued, would let the agency use AI tools without raising the risk of personal data leaks. He also pointed to persistent data silos — where information is fragmented across ministries and departments — as a barrier. “To move into the AI era, we must break down these walls,” he said, noting that some regulations currently block data sharing at the source and will need revision.
Describing the operational goal for stronger AI capabilities, Lee said the agency wants AI to help the police better protect public safety. He warned that AI-enabled crimes are already emerging below the surface and that some offenders are automating operations with AI — an imminent threat. AI-facilitated crime is evolving into CaaS (Crime as a Service) and advancing quickly; if the police lag, the public will bear the consequences. “Police must adopt AI proactively and faster than criminals,” Lee said. “That’s why we started this work early and are moving quickly.”
Most Commented