예측치안에 대한 법적 규율 - 인공지능 위험평가에 기반한 예측치안 기술의 도입현황과 입법과제

• 영문명: Legal Regulation of Predictive Policing : Current Implementation of Predictive Policing Technologies Based on AI Risk Assessment and Legislative Challenges
• 발행기관: 충북대학교 법학연구소
• 저자명: 유승익(Seung Ik Yoo)
• 간행물 정보: 『과학기술과 법』제15권 제2호, 95~122쪽, 전체 28쪽
• 주제분류: 법학 > 법학
• 파일형태: PDF
• 발행일자: 2024.12.31

국문 초록

This study examines the adoption and application of AI-based predictive policing technologies while addressing legislative challenges to mitigate their risks. Predictive policing technologies have revolutionized public safety and crime prevention by enabling advanced data analysis and algorithmic decision-making. However, they also pose significant challenges to constitutional values and fundamental rights. This paper focuses on the potential erosion of key constitutional principles, including the presumption of innocence, due process, and the principle of accountability, in the context of machine learning and data-driven decision-making. It highlights the dual potential of predictive policing technologies to be extensively applied across various stages of criminal justice—from crime prediction during police operations to investigations, bail, sentencing, and corrections—while risking infringements on individual freedoms, privacy, and reinforcing systemic biases. To address these issues, the paper underscores the necessity of clear legal frameworks regulating predictive policing technologies and proposes legislative directions that establish explicit boundaries between permissible and prohibited uses. Key recommendations include the development of criteria that differentiate acceptable and restricted uses based on technology purpose and operational environment, strengthening oversight mechanisms for high-risk AI systems, and implementing mandatory measures such as robust data governance, risk management protocols, and operational logging. Furthermore, recognizing the cross-border nature of AI technology, the paper emphasizes the importance of international cooperation to achieve regulatory interoperability and prevent legal conflicts and regulatory arbitrage resulting from divergent national regulations. Strengthening institutional mechanisms to ensure redress and legal protections for individuals affected by predictive policing technologies is another critical focus, aiming to safeguard fundamental rights through post-incident remedies and recovery procedures. In conclusion, the paper argues for the establishment of a cautious and systematic regulatory framework that balances the transformative potential of predictive policing technologies with the protection of constitutional values and fundamental rights. Regulations should function not as barriers to technological innovation but as essential safeguards to prevent misuse and ensure equitable and safe deployment. Policymakers and regulatory authorities must adopt a balanced approach prioritizing long-term societal impacts over short-term efficiency, thereby creating sustainable legal and social systems to support the responsible use of predictive policing technologies.

목차

Ⅰ. 문제제기

Ⅱ. 인공지능 위험평가 기반 예측 치안의 개념과 현황

Ⅲ. 예측치안에 대한 법적 규율의 필요성

Ⅳ. 외국의 입법동향과 입법과제

Ⅴ. 결론

키워드

인공지능 위험평가 예측치안 프리카스 유럽연합 인공지능법 고위험 인공지능 Artificial Intelligence Risk Assessment Predictive Policing Pre-cas European Union Artificial Intelligence Act High-Risk AI