정책결정 과정에서 정확한 정보와 엄밀한 분석의 중요도가 높아지고 있다. 정책 수립 과정에서 이해관계자를 설득하고 정책의 실패 가능성을 최소화하기 위해 통계를 활용하는 것은 일상화되었다. 하지만 무비판적인 통계의 제시나 착시적 통계 수치의 악용 등은 정책의 효과성을 반감시키고, 극단적으로는 의도한 것과 정반대의 효과를 초래하기도 한다.

2000년을 전후하여 재정사업 성과관리제도가 도입되면서 우리나라는 정책과 관련된 각종 데이터를 수집관리하였고 정책의 효과를 측정하는 등 정책의 과학화를 위한 노력을 시도하였다. 하지만 새로운 정책 수립 시 과거 정책으로부터 획득한 증거를 이용하거나 정책실험을 통해 효과성을 검증하는 등 적극적 의미의 증거기반정책 수립은 아직까지 국가 정책 전반으로 확산되지 못하고 있다.

본 연구는 점차 복잡해지는 사회구조하에서 보다 효과적이고 시의적절한 정책결정을 도출해 내기 위한 인공지능 기반 정부 지원 통합체계의 구축 가능성을 검토한다. 이는 단순히 고용량고연산 기술을 정부부문에 도입하는 것을 의미하는 것이 아니다. 공공부문의 업무, 특히 여러 돌발 상황을 염두에 두어야 하는 정책의 수립 및 집행 과정에 인공지능기술을 도입하기 위해서는 기존의 업무 체계 및 인력에 대한 대대적인 정비가 필요하다. 본 연구는 정부부문의 특성을 고려할 때 현시점에서 인공지능기술 도입의 효용과 장애요인을 분석하고 향후 인공지능기술 도입의 확대를 위해 선결되어야 할 정책 과제를 도출한다.

Artificial intelligence (AI) or big data analysis are increasingly applied in very diverse fields such as marketing, healthcare, management, manufacturing, economic policy, not to mention professional sports. As available data explodes and AI technologies advance fast, the combination of these two is already transforming businesses and our lives. The goal is to augment human intelligence to be more accurate in detection, prediction, or decision-making. AI technologies offer new possibilities for policy-making and public administration helping the governments become more data-driven. In this respect, the combination of AI technology and government policy is an inevitable option to greatly increase the efficiency and effectiveness of the public service. At this point, the discussions for the successful introduction of AI technology in the public sector and policy-making are very urgent.

In this study, we explore the possible benefits of AI technology in the public sector by clarifying the connections and distinctions between AI technology and the existing evidence-based policy-making process. Besides, we identify the factors that hinder the introduction of the technology and suggest policy recommendations to overcome the challenges. We intend to provide information on how to build a government system for data-driven decision-making.

Chapter 2 argues that to reap the benefits of AI, it is necessary to change the current policy-making process to be more evidence-based. A fundamental transformation must be accompanied in the government system by reviewing the overall decision-making process including practices in collecting, managing, and sharing data. This chapter studies the development and application of evidence-based policies in Korea and other major countries and discusses the role of AI technology. Chapter 2 also deals with the possibilities and constraints when implementing AI technology in public policies especially in selecting policy targets.

Chapter 3 empirically presents the benefit of using AI technology in business support policy. It presents methods and results of applying machine learning, an AI technology, to the actual Korean government’s business support policies. We discuss important issues in the actual implementation of machine learning in policy targeting. The Korean government is implementing hundreds of business support policies to enhance corporate competitiveness and strengthen the overall competitiveness of the industry. The success of a business support policy can largely depend on the selection of a company that meets the purpose of support. This chapter discusses ways to improve the effectiveness of government support programs by comparing the results of selected targets by machine learning with the actual recipients of the programs.

The biggest advantage of AI technology lies in its infinite scalability. The advantage can take place when policy-makers use relevant data and the results from the analysis to improve the policy. The government holds a vast amount of personal and corporate information through administrative data. When the information is combined with private big data, the advantage becomes even greater. Besides the business support policy, AI technology can significantly increase policy effectiveness when it is applied to education, medical care, policing and defense to name a few. Still, we face the gap between the possibility and reality in the AI application. In Chapter 4, we conducted a survey on the perception and actual status of the introduction of artificial intelligence into the policy-making process with public sector workers related to business support policy and business people who are beneficiaries of the policy. The survey results showed that the respondents were recognizing the positive aspect of the application of AI in terms of enhancing fairness and reliability in selecting the policy targets. We identified technical, institutional, and cultural implementation barriers through the survey and suggested recommendations to overcome the barriers and increase the effectiveness of business support policies. The policy recommendations include a suggestion to form an organization that can oversee data management and artificial intelligence-related tasks scattered across various ministries. The organization is necessary to reinforce the strategic management capacity of information assets at the national level.

In conclusion, chapter 5 suggests that the current Korean national strategy for AI should include a data-driven government strategy across the public sector. The data-driven government strategy should prioritize activities in transforming governments’ decision-making process to be data-driven or evidence-based. The chapter ends with a four-step guide to a data-driven organization.

발간사
요 약

제1장 서 론
　참고문헌

제2장 증거기반정책의 수립과 인공지능기술
　제1절 증거기반정책의 수립
　제2절 인공지능기술과 공공서비스
　제3절 인공지능기술 활용 데이터 기반 정책결정
　참고문헌

제3장 AI 기반 기업 지원정책
　제1절 기업 지원정책에의 AI 기술 적용
　제2절 중소기업 자금 지원사업에의 AI 기술 적용
　제3절 지원사업에의 AI 기술 적용 결과
　제4절 소결 및 시사점
　참고문헌

제4장 AI 기반 기업 지원정책 도입의 장애요인과 극복방안
　제1절 기업 지원정책 추진 시 인공지능(AI) 도입에 대한 인식 및 실태조사
　제2절 기업 지원정책 부문의 AI 기술 활용 영향요인
　제3절 시사점 및 정책제언
　참고문헌

제5장 맺음말
　참고문헌

부 록
ABSTRACT