대학의 교양수학에서 인공지능과 수학 관련 교과목 개발에 관한 연구

최재길,심상길 한국교양교육학회 2024 교양교육연구 Vol.18 No.4

본 연구는 인공지능과 수학 관련 교과목 개발을 위해 수학과 교육과정, 인공지능과 수학 관련 연구 등을 분석하여 학습요소 및 내용을 선정하고, 대학의 교양수학에서 인공지능과 수학 관련 교과목 구성 및 운영에 대한 방안을 제시하였다. 인공지능의 원리를 수학에서 다루기 위해 필요한 내용은 주로 미적분학, 확률과 통계, 선형대수학에 관련된 내용으로, 이와 같은 수학 학습 내용은 인공지능의 기능 중에서 데이터 분석 및 표현, 신경망, 클러스터링, 비지도학습, 최적화 등을 위해 필요하므로 대학교양수학의 인공지능과 수학에 관련 교과목에서는 이러한 내용이 필수적으로 포함되어야 한다. 또한, 인공지능과 수학 관련 교과목은 머신러닝을 이해하기 위한 기초수학으로 필수적인 선형대수학과 다변수 미적분학의 여러 주제들을 한 학기에 배우기 어려우므로 본 연구에서는 인공지능과 수학 관련 교과목을 <인공지능 수학1>과 <인공지능 수학2>의 두 과목으로 나누어 대학생들이 자신의 능력과 필요에 따라 선택하여 수강할 수 있도록 제안한다. 마지막으로, 인공지능과 수학 관련 교과목은 인공지능의 원리를 이해하기 위한 이론 중심의 수업도 중요하지만 널리 쓰이는 프로그램 언어인 파이썬에 대한 기초문법을 학습하고, 이를 이용한 데이터 분석과 자료정리 및 전처리에 대한 학습을 통해 실생활에서 찾을 수 있는 데이터 분석과 인공지능, 딥러닝 프로젝트 등을 적용하는 응용학습에 대한 연구도 필요하다. In order to develop courses related to artificial intelligence and mathematics, this study provides plans for the compositionand operation of courses related to artificial intelligence and mathematics in college general mathematics by analyzing themathematics curriculum, artificial intelligence and mathematics-related research, and selecting learning elements and contents. The content needed to deal with the principles of artificial intelligence in mathematics is mainly related to calculus, probabilityand statistics, and linear algebra. Such mathematical learning content is one of the functions of artificial intelligence. Sinceit is necessary for analysis and expression, neural networks, clustering, unsupervised learning, and optimization, these contentsmust be included in artificial intelligence and mathematics-related courses in college general mathematics. In addition, incourses related to artificial intelligence and mathematics, it is difficult for students to learn various topics such as linearalgebra and multi-variable calculus, which are essential as basic mathematics for understanding machine learning, all in onesemester. Thus, in this study, I propose that we divide the course into two subjects, <Artificial Intelligence Mathematics1> and <Artificial Intelligence Mathematics 2>, so that college students can select and take the courses according to theirabilities and needs. Lastly, while theory-oriented classes are essential for understanding the principles of artificial intelligencein artificial intelligence and mathematics-related courses, applied learning is also crucial. This includes learning the basicsof Python, a widely used programming language, for data analysis, data organization, and preprocessing. Additionally, applyingthese skills in real-life projects involving data analysis, artificial intelligence, and deep learning is necessary for comprehensiveeducation.

인공지능과 민주주의에 관한 소고

윤수정 헌법재판연구원 2021 헌법재판연구 Vol.8 No.2

In the era of Phono Sapiens, where everything is solved with a smartphone, artificial intelligence (AI) is closely permeating our lives. Now, the problem of artificial intelligence is not a problem of “if,” but a realistic problem that has to deal with “how fast” and “to what extent” to allow, and is developing into an important problem that can determine the daily life of the national community and even the future of mankind. As artificial intelligence advances into various areas of human life, another notable area is politics. As politics became digital, the political subject changed to a large number of citizens, various political agendas were raised, and the way of political participation changed. The development of artificial intelligence also poses various opportunities and risks in relation to democracy. Democracy, a constitutional principle that seeks to realize the thoughts of individual freedom and self-determination within the national community, is essentially realized through representative democracy. However, in reality, representative democracy faced a crisis due to the distortion of democratic representation and accountability, and the continued COVID-19 situation accelerated this. Artificial intelligence is being used to overcome the limitations of representative democracy, but its influence is like a double-edged sword. Although positive effects such as strengthening universal information access, easing representative democratic limits, building artificial intelligence governance, expanding the realm of human freedom, and increasing political fairness are expected, there are side effects such as the emergence of a new surveillance society, distortion and insufficiency of political opinion. Therefore, in order for artificial intelligence to coexist with democracy, a normative response aimed at minimizing side effects from a democracy perspective is necessary. The basic principle of this normative response should be ‘Human-Centered Artificial Intelligence’. Human dignity should be respected in all stages of artificial intelligence design, development, and utilization, and artificial intelligence should only be used as a means of supplementing human autonomy. It is time to actively think about what to do with artificial intelligence rather than what can be done with artificial intelligence to realize democracy. 스마트폰 하나면 모든 것이 해결되는 포노 사피엔스(Phono Sapiens)의 시대, 인공지능(AI)은 우리의삶에 밀접하게 스며들고 있다. 이제 인공지능의 문제는 ‘만약(if)’의 문제가 아닌 ‘얼마나 신속하게(how fast)’, 그리고 ‘어느 범위까지(to what extent)’ 허용할 것인가를 다루어야 하는 현실적인 문제이며, 국가공동체의 일상, 나아가 인류의 미래를 좌우할 수 있는 중요한 문제로 발전하고 있다. 인공지능이 인류의 삶의 다양한 영역으로 진출하면서 또 하나 주목할 만한 영역은 정치이다. 정치가디지털화되면서 다수의 국민으로 정치 주체가 변화하였고, 다양한 정치 의제가 제기되었으며, 정치 참여의 방식이 변화되었다. 이러한 디지털 정치는 이제 인공지능을 정치에 활용하는 것으로 발전하고 있다. 인공지능의 발전은 민주주의와의 관계에서도 다양한 기회와 위험을 내포하고 있다. 개인의 자유와 자기결정의 사고를 국가공동체 내에서 실현하고자 하는 헌법원리인 민주주의는 본질적으로 대의민주주의를 통해 실현된다. 그러나 현실에서 대의민주주의는 민주적 대표성과 책임성의 왜곡이발생하면서 위기에 직면하였고, 지속된 코로나19 상황은 이를 가속화시켰다. 이러한 대의민주주의의 한계를 극복하기 위해 인공지능이 활용되고 있지만, 그 영향은 양날의 칼과 같다. 보편적인 정보접근성이강화, 대의민주주의 한계의 완화, 인공지능 거버넌스 구축, 인간의 자유 영역 확대, 정치의 공정성 증대등의 긍정적 효과가 기대되기도 하지만, 새로운 감시 사회의 출현, 정치적 의사형성의 왜곡과 불충분성초래, 소수 디지털 기업에의 권력 집중 등의 부작용도 간과(看過)할 수만은 없다. 그러므로 인공지능이민주주의와 공존(共存)하기 위해서는 민주주의 관점에서 유익한 목적의 인공지능의 개발은 장려하면서도 부정한 목적의 개발과 이용, 그리고 부작용을 최소화하는 것을 목표로 하는 규범적 대응이 필요하다. 그리고 이러한 규범적 대응의 기본원칙은 ‘인간 중심의 인공지능’이 되어야 한다. 인공지능의 설계, 개발 및 활용의 전 단계에서 인간의 존엄성을 존중하여야 하며, 인공지능은 인간의 자율성을 보완하는 수단으로서만 활용되어야 한다. 이제 민주주의 실현을 위해 인공지능으로 무엇을 할 수 있는지보다 인공지능으로 무엇을 해야 하는지를 적극적으로 고민해야 할 시점이다.

인공지능에 의한 발명의 발명자성에 대한 소고

이상호 한국지식재산학회 2022 産業財産權 Vol.- No.71

Under Korean patent laws, a patent right is initially vested in an inventor, and the justification of the patent right needs to be reviewed under the consideration of whether the recipient can be qualified as an inventor. This paper reviews three philosophical justifications for patent rights and applies them to artificial-intelligence-generated inventions and concludes that an inventor should be a natural person, and artificial intelligence's inventorship is not justified. First, under Locke's labor theory, every human being has property rights to his own person from birth and can expand his ownership, including patent rights, through his manual or mental labor. Artificial intelligence is a man-made object and does not have any basis for expanding property rights therefrom through labor as it is bereft of any innate rights whatsoever to own itself. Under Hegel's personality theory, a natural person has "the right to direct his will upon any object" and is a bearer of free will, which also can be a basis of the right to a patent. Free will is not merely the act of calculation, it's a reflection of Being itself. Artificial intelligence cannot have such self-purpose and dignity as human beings have and thus cannot be justified as an inventor. Under Bentham's Utilitarian view, it is presumed that an individual responds to incentives, and that this response benefits society. However, artificial intelligence lacks any motivation to devise an invention seeking a monopoly. One might argue that if artificial intelligence's inventorship is denied and no natural persons are qualified as an inventor and no one would be able to reap the benefits arising from the invention generated by artificial intelligence. It might be concerning as this will diminish scientists' motivation to develop the technology and eventually lead to a failure to maximize the utility in society as a whole. However, scientists' efforts should be protected only to the extent appropriate to their contributions. No one should be granted the rights to an invention that he or she has not contributed to. Otherwise, once the artificial intelligence technology reaches AGI (artificial general intelligence) or ASI (artificial superintelligence), only small number of owners would enjoy the limitless fruits of such artificial intelligence. Additionally, once artificial intelligence gains legal status, it can expand its rights boundlessly in society, which would pose a menace to humankind. In these cases, the inventorship of artificial intelligence should be denied under the Utilitarian view as it fails to maximize its utility in society. Accordingly, it is submitted that only natural persons are eligible for an inventor. If artificial intelligence generates an invention with no human contribution, the invention becomes the property of the public. However, if a natural person's contribution is insufficient to satisfy the inventorship requirement, it may be necessary to expand the scope of inventorship to include partial contributions from natural persons. In this regard, the work for hire doctrine under copyright provides a hint on how to enlarge the boundary of inventorship. It is suggested that a natural person employer who defines a problem and plans to find the solution using artificial intelligence, as a creative employee, should be credited as an inventor under the ‘invention for hire’ doctrine even if his contribution to a conception of an invention does not meet the current inventorship requirements. 특허법에 따르면, 특허권은 원시적으로 발명자에게 주어지므로 특허권의 정당성을 논함에 있어 권리를 받는 자가 발명자로 인정되어 권리를 누리는 것이 합당한지 검토해야 할 필요가 있다. 본 고는 특허권에 대한 세 가지 철학적 정당성을 검토하고 인공지능에 의한 발명에 적용하여 오직 자연인만 발명자로 인정받을 수 있으며 인공지능의 발명자성은 정당화되지 못한다고 결론 내렸다. 먼저, 로크의 노동이론에 따르면 모든 인간은 생득적으로 자기 자신에 대한 재산권적 권리를 가지며 특허권을 포함해 자신의 정신적 육체적 노동을 통해 소유권을 확장해 나갈 수 있다. 하지만 인공지능은 사람이 만든 객체에 불과하여 자기 자신으로부터 노동을 통해 확장할 수 있는 어떤 내재적 권리의 기반을 가지고 있지 못하다. 헤겔의 인격 이론에 따르면 자연인은 자신의 의지를 객체에 투영할 수 있는 권리를 가지며 특허권의 근거가 되는 자유의지의 담지자이기도 하다. 자유의지는 단순한 정보의 선택을 의미하지 않으며 존재 자체로부터 나오는 자기 자신의 목적의 투영이다. 인공지능은 인간과 같은 자기 목적성이나 존엄성을 가지고 있지 못하므로 발명자로 정당화될 수 없다. 벤담의 공리주의에 따르면, 인간은 유인에 반응하는 존재로 상정되며 이러한 인간의 반응이 사회의 유용을 증대시키는 역할을 한다. 하지만, 인공지능은 이러한 유인에 반응하여 발명의 독점권을 추구하는 존재가 아니다. 혹자는 만약 인공지능의 발명자성이 부정되고 발명자로 인정될 수 있는 자연인이 존재하지 않는 경우, 누구도 인공지능이 발명한 결과물에 대한 유익을 취할 수 없게 될 것이고, 이는 결국 인공지능을 연구하는 과학자들의 사기를 저하해 사회 전체 유익이 최대화되지 못한다는 위험을 불러올 수 있다고 염려할 수 있다. 하지만 과학자들의 노력은 그에 상응하는 정도의 보호만 받을 필요가 있다. 누구도 자신이 발명하지 않은 것에 대한 권리를 주장할 수는 없다. 인공지능기술이 초인공지능 내지는 강인공지능에 이르게 되면 소수의 사람이 인공지능이 개발한 열매를 누릴 수 있는 위험이 있다. 또한, 인공지능이 자신의 법률적 지위를 가지게 되면 특허권은 자신의 발명에 대한 권리를 사회에서 무한대로 확장할 수 있는 근거가 될 수가 있고, 이는 인류가 위협될 수 있다. 이 두 경우 모두 인공지능을 발명자로 인정하는 것은 부정되어야 하는데 이는 공리주의적 관점에서 사회 전체적 유용성을 최대화하지 못하기 때문이다. 따라서, 자연인만 발명자로 인정받아야 하며, 만약 인간의 개입 없이 인공지능에 의한 발명이 이루어지면 이는 결국 공공재로 취급되어야 한다. 그러나 자연인이 인공지능을 사용하며 발명에 이르는 과정을 기획한 경우의 비록 발명의 착상에 대한 발명자의 기여가 비록 발명자 요건을 만족시키는 정도는 아니라 하더라도 발명자성의 범위를 보다 확대하여 발명자성을 부여하는 것이 합리적이라 하겠다. 이와 관련해서 저작권법의 업무상 저작물 제도는 발명자성의 외연을 넓히는 단초를 제공한다. 본 고는 자연인 사용자가 문제를 정의하고 해결책을 찾기 위해 인공지능을 창의적 근로자로 사용한 경우 해당 자연인의 발명 착상에 대한 기여가 비록 특허법에서 요구하는 발명자성에 미치지 못한다고 하더라도 업무상 발명(invention for hire)제도를 통해 발명자로 인정해줄 것을 제안한다.

인공지능 예술에 담긴 커뮤니케이션에 대한 고찰

한지윤,최유미 (사)한국만화웹툰학회 2025 만화웹툰연구 Vol.5 No.2

Artificial intelligence has been a great inspiration for artists. Artificial intelligence art was created in various ways using artificial intelligence as a creative tool. Since then, there has been a dispute over the possibility of art subjects. In academia, many studies have been conducted on this subject. However, most of the research contents are mainly based on the morphological analysis of the completed creative results. So I felt the need for research in a different approach. Artificial intelligence and art were the same research subjects. And I want to find the intersection between the two and explore artificial intelligence art. First, understand the meaning of communication. The whole system of society consists of communication. Therefore, communication will be found in artificial intelligence and art. Second, research is conducted on the development of artificial intelligence and the learning structure of artificial intelligence and human intelligence. The process of delivering data to learn intelligence is analyzed with communication theory. Third, artificial intelligence works of art are analyzed based on the research conducted. Finally, I organize about communication. It was confirmed that each has communication in artificial intelligence, art, and artificial intelligence art. Communication is complex communication from 'double coincidence' to 'virtual double coincidence' depending on the sender and receiver. In the process of information-notification-understanding communication, the screening subject plays an important role. Artificial intelligence art will be understood as humans have selected the first information and notifications, and eventually humans will finally select and present their works. Therefore, it is argued that artificial intelligence art is the result of human design. However, when artificial intelligence communicates on its own, it has an abstraction that has coincidence that humans do not intend. Artificial intelligence art is created through communication to make it original with algorithms. Repeated considerations of the coincidences of art/non-art will lead to the evolution of art. The virtual double coincidence in artificial intelligence communication has discovered the peculiarities of artificial intelligence art. Finally, in the paper, the 'abstractness' of 'accidental abstraction' refers to a characteristic that is constantly changing without being fixed from a physical perspective. This is expected to affect visual expression from an art perspective, and research is needed.

AI 창작물의 저작권 보호는 필요한가

박성준,김상겸 연세대학교 법학연구원 2024 법학연구 Vol.34 No.2

It is the era of artificial intelligence. Technological innovation, which has been a constant driving force for the growth of mankind, has emerged a new type of technology called artificial intelligence. With the development of digital and information and communication technologies, artificial intelligence technology has opened the horizon of the new era of the 4th industrial revolution and provided many conveniences to mankind. Meanwhile, artificial intelligence has taken human life to the next level and presented humanity with numerous legal challenges. The problem of artificial intelligence creations is also a representative example of this. Artificial intelligence, which has been developing over and over, has expanded its scope to the realm of creation, which was said to be the exclusive property of mankind, and has caused problems with copyright and responsibility, bringing about a great impact on the human intellectual property system. Looking at this situation, this paper was written to find a way to achieve harmony with the human intellectual property system by regulating and protecting artificial intelligence creations within the copyright field, and the results of the study are as follows. First, although various artificial intelligence creations are produced, it was confirmed that artificial intelligence creations are in a legal blank state under the current intellectual property system. Since it is not recognized as a work under the copyright law, it is not possible to regulate them within the current legislation governing human creation. As a result of examining their need for protection, it may be contradictory, but it was concluded that artificial intelligence creations need to be regulated within the current legal system for human creation. Even considering the purpose of the Copyright Act, which is the protection of human creations and the development of related industries, the necessity of regulation and protection can be recognized. However, in order to protect artificial intelligence creations, numerous issues of copyright law, such as the issue of copyrightability of creations, copyright attribution, and protection method, must be decided in the current legislation. Considering that artificial intelligence cannot be a subject of rights and obligations at present and the current technological ability to create creations without human intervention, the subject of rights of creations should be human. Based on this, the constitutional approach to the knowledge and information society as a protection measure, “introduction of electronic personality” to prepare for the upcoming era of strong artificial intelligence, legal improvements such as “Amendment of Copyright Act” and “Enactment of Special Act on Artificial Intelligence” as a way to realize the recognition of copyright and the establishment of authors, and the production of “Ethical Guidelines” as a concrete system, the registration system as a technical indication obligation, and external measures were proposed. However, since these measures also attempt a legal approach to artificial intelligence from the beginning, social consensus should be reached through further discussion on their concrete realization measures. Regardless of the field, artificial intelligence will create new legal issues in a completely different way than before. Now, we must start an in-depth discussion to present a new paradigm to respond to new issues in artificial intelligence.

인공지능을 활용하여 창출된 발명에 대한 바람직한 진보성 법리

정차호,김유진,이서영 충남대학교 법학연구소 2023 법학연구 Vol.34 No.1

Nowadays, with the development of artificial intelligence technology, many inventors are utilizing artificial intelligence to create their inventions. Due to the changed research environment, a change or at least an assessment of current inventive step (non-obviousness) jurisprudence is required. In other words, it is necessary to review whether current inventive step (non-obviousness) jurisprudence, which is summarized as “a person having ordinary skills in the art (PHOSITA)” and “obvious extraction”, is still well grounded in the era of artificial intelligence. Some have argued that wide utilization of artificial intelligence makes invention easier and increases the concern over too many patents. This paper analyzes academic discussions from Korea, China, Japan and the United States on new inventive step jurisprudence which should be applied to inventions created utilizing artificial intelligence (assuming weak artificial intelligence). The discussions are mainly divided into (1) the argument that the current concept of “PHOSITA” should be changed to the concept of “ordinary artificial intelligence”, and (2) the argument that the utilization of artificial intelligence should be described in the patent specification. Regarding the argument (1) above, this paper proclaims that a weak artificial intelligence cannot be a (co-) inventor and is only a more powerful tool. Thus, inventive step should be judged based on the standard of “smarter PHOSITA” who can utilize artificial intelligence. Regarding the above argument (2), because artificial intelligence may be a trade secret and it is practically difficult to impose a penalty on a specification that does not describe utilization of artificial intelligence, the paper disagrees with the argument. Although utilization of artificial intelligence will make invention easier and increase the number of patent applications, the threshold of inventive step will also be heightened due to the standard of “smarter PHOSITA”. As a result, the number of patent registrations is not likely to increase significantly. In conclusion, this paper believes that the basic framework of the current inventive step (non-obviousness) jurisprudence can be maintained in the era of weak artificial intelligence.

중등학생 대상 인공지능 윤리역량 평가척도 개발

배진아,이정훈,홍미선,조정원 한국컴퓨터교육학회 2022 컴퓨터교육학회 논문지 Vol.25 No.6

Artificial intelligence technology is a representative future industrial technology that has a great impact on society and the economy as a whole. However, as artificial intelligence becomes closer to our lives, unexpected ethical issues are also emerging. Accordingly, major countries around the world are making various efforts, such as preparing ethical standards to secure trust in artificial intelligence. In addition, the need for artificial intelligence ethics education for students who will become users and developers of artificial intelligence has emerged. However, there are not enough research to establish ethical standards for artificial intelligence ethics education and strengthening of executive power. Therefore, the purpose of this research is to develop a scale that can measure the ethical competence of artificial intelligence for middle and high school students. This research was conducted with the aim of developing a process evaluation tool to foster artificial intelligence ethics and strengthening of executive power of artificial intelligence ethics. To this end, the items were modified and supplemented through scale item analysis and expert advice previously studied, and data were collected by applying them to educational programs. After that, through this survey, the artificial intelligence ethical competency scale was developed through stages such as validity and reliability test of variables and confirmation factor analysis. Through the artificial intelligence ethical competency evaluation scale developed through this study, it will be possible to comprehensively evaluate the artificial intelligence ethical literacy and execution ability of middle and high school students. 인공지능 기술은 사회, 경제 전반에 광범위한 영향을 미치는 대표적인 미래 산업기술이다. 그러나 인공지능이 우리 삶에 밀접해지면서 예상치 못한 윤리적 이슈도 함께 나타나고 있다. 이에 주요 국가들은 신뢰 확보를 위한 인공지능 윤리기준을 마련하는 등 다양한 노력을 기울이고 있다. 또한 인공지능의 이용자이면서 개발자가 될 학생들을 위한 인공지능 윤리 교육의 필요성이 대두되었다. 하지만, 인공지능 윤리교육을 위한 윤리기준을 마련하고, 윤리에 대한 실천력을 강화를 위한 연구는 충분치 않다. 이에 본 연구에서는 중 ·고등 학생을 대상으로 인공지능 윤리역량을 측정할 수 있는 척도를 개발하는 것을 목적으로 한다. 인공지능 윤리소양 함양과 인공지능 윤리 실천력을 강화하기 위한 수행형 기반 측정 도구 개발을 목표로 연구하였다. 이를 위해 기존에 연구된 척도 문항 분석과 전문가 자문을 거쳐 문항을 수정 및 보완하고, 교육 프로그램에 적용하여 데이터를 수집하였다. 이후 본조사를 통해 변수의 타당성 및 신뢰도 검정, 확인적 요인분석 등의 단계를 거쳐 인공지능 윤리역량 평가척도를 개발하였다. 본 연구를 통해 개발된 인공지능 윤리역량 평가척도를 통해 중 ·고등학생의 인공지능 윤리 소양과 윤리 실천력 등을 종합적으로 평가할 수 있을 것이다.

인공지능에 의한 사이보그 형 자동화 의사결정에 대한 고찰

양종모 충북대학교 법학연구소 2021 과학기술과 법 Vol.12 No.1

Artificial intelligence research started with computers trying to take the place of what humans have to do intelligently. Research started in the 1950s went through several ups and downs and reached the same period of the third boom as it is now. With the shocking situation where Alpha Go defeated Lee Se-dol, the strongest player in Go, in 2016, interest in artificial intelligence increased rapidly, and stories about the future situations that artificial intelligence will bring about, such as the possibility and singularity of artificial intelligence, are talked about even among the general public. Certainly, artificial intelligence has the potential to fundamentally change the way humans live, not just technological advances. The real value of artificial intelligence algorithms is prediction. At the center of all the algorithmic revolutions that have taken place in the various fields of modern society, there is one unchanging purpose: prediction. The algorithm for predicting the risk of recidivism used in US law enforcement fits precisely with the principle that individual humans behave in a predictable and consistent manner. While using the algorithm of predicting the risk of a second offender, an artificial intelligence algorithm, not a judge, actually takes the place of the defendant's decision on whether or not to be arrested or sentenced. Automated decision making continues to grow at an unprecedented rate and benefits because machine learning offers the possibility to extend the automated decision-making process, allowing wider, deeper decisions without human intervention. It is also very large and is expected to continue to expand its use. However, the risks are as great as the benefits of artificial intelligence algorithms. The risks of artificial intelligence algorithms outweigh the individual risks of any algorithmic technology specifically applied. Despite the widespread use of artificial intelligence in the social sphere, the reality is that the problems of artificial intelligence cannot be properly grasped. However, people who were treated differently due to automatic decision-making by artificial intelligence began to question the fairness or fairness of artificial intelligence, and in particular, while causing sensitive problems such as discriminatory results, controversy surrounding artificial intelligence algorithms gradually became more and more controversial. It is expanding. Even if a decision made by artificial intelligence is wrong or causes negative effects such as discrimination, there is a big problem that there are few means to object to this. In a situation where automated decision-making is more likely to threaten human autonomy and dignity, in order to secure an individual's autonomous personality, make sure that the entire process of the individual's automatic decision-making that affects him or her is fair and error-free. It is of utmost importance to know and allow humans to directly intervene if necessary. Although available data is increasing and algorithms are gradually improving, artificial intelligence algorithms do not completely eliminate the uncertainty posed by decision making. Eventually, there are cases in which humans, as finalists, have to make important decisions in situations where uncertainty has not been resolved. It is necessary to build a system in which artificial intelligence algorithms and human experts cooperate well to make the best decisions, and in addition to the direction of such a collaboration system establishment, the problem of legal responsibility due to errors in decision-making in artificial intelligence algorithms or collaboration systems is necessary. I want to deal with it in this paper.

인공지능에 의한 행정상 자동결정에 대한 규율과 권리구제

김휘식 경북대학교 IT와 법연구소 2021 IT와 법 연구 Vol.0 No.22

Artificial intelligence, one of the core technologies leading the 4th industrial revolution, is expected to bring about an innovation in administration. In the past, e-government has implemented various public services through Internet and brought innovation in administrative affairs. Intelligent informatization in administration takes one step more to make automatic administrative decisions by using artificial intelligence in various fields of administration. In this paper, we examine the aspects of regulations and remedy of rights in automatic administrative decisions with artificial intelligence. In the case of automatic administrative decisions with artificial intelligence, regulations and remedy of rights may vary depending on various types of administrative actions, such as administrative disposition, public contracts, realakte, and administrative rules. When establishing the the regulations of automatic administrative decisions with artificial intelligence, lawmakers should consider rule of law, administrative procedures, opacity and bias of machine learning algorithms, and personal information protection. When making administrative disposition with artificial intelligence, it is reasonable to allow discretionary actions as well as binding actions considering the nature of AI. In order to enforce public contracts and realakte with artificial intelligence, legal grounds will be required. Automatic administrative rules with artificial intelligence will be allowed only when there is social consensus based on democratic legitimacy. When referring to overseas regulatory trends, it is necessary to recognize the right to object against automatic administrative decisions and to grant the right to request an explanation of the AI′s algorithms. On the other hand, it is necessary to keep in mind the principles of science and technology regulation, such as so-called negative regulation-a method of specifically enumerating the rights limited and permitting other matters in principle. Regarding the remedy of rights, the current law does not reflect the automatic administrative decision with artificial intelligence, so remedy of rights is insufficient. In relation to state compensation liability, if public officials using artificial intelligence inflict damage on other persons, the negligence of public officials should be presumed. If damage occurs in public structures due to artificial intelligence errors, defects should be presumed when the probability of the defects is proved. When it comes to an appeal suit, automatic administrative decisions with artificial intelligence-for example, machine learning algorithms that reflect specific disposition criteria- should be recognized as administrative disposition. With regard to the illegality of administrative disposition, the issues are rule of law, administrative procedures, opacity and bias of machine learning algorithms, personal information protection, etc. In addition, a party suit should be used for state compensation claims or public contracts.

인공지능(AI) 시대의 도래와 법제도의 방향에 관한 논의 - 독일의 인공지능 정책을 중심으로 -

송기복 한국경찰법학회 2020 경찰법연구 Vol.18 No.2

Artificial intelligence is gradually moving away from the field of research and is being used throughout everyday life, including self-driving, chatbot, biometric recognition, smart home, and medical devices. In addition, the utilization and application of big data is affecting various fields such as society, economy, politics, and law. Already in major advanced countries, discussions have been held to adjust the current legal system for the protection of data protection and privacy in telecommunications, industry, medical and financial sectors, the risk of discrimination by artificial intelligence, the issue of responsibility for artificial intelligence systems, and the establishment of ethical and artificial intelligence policy systems for artificial intelligence. Meanwhile, Germany has developed a high level of artificial intelligence research infrastructure (Infra) in connection with the federal government, state governments, research institutes, industries, and universities, and has developed widely in social, economic, and industrial sectors to make useful use of the benefits and convenience that artificial intelligence technology will bring. On the other hand, the legal system is being reviewed to cope with changes that will occur in various fields such as society, ethics, legal system, and culture due to the emergence of artificial intelligence technology and to prepare for the risks posed by artificial intelligence. First of all, based on the ethics needed in a series of processes such as designing, manufacturing, and utilizing artificial intelligence through technical standardization, and academic dialogue such as social, economic, and cultural sectors, the expansion of convergence on artificial intelligence's social impact, future effects by technology, and discussions on designing artificial intelligence to prevent artificial intelligence from threatening human safety. In order to establish a legal system that can utilize artificial intelligence technology without conflict with existing social values and legal systems, this school intended to provide a forum for discussion on the direction of artificial intelligence policy in Korea, although it is insufficient. 인공지능은 점차 연구의 장(場)에서 벗어나 자율주행, 채팅봇, 생체 인식, 스마트 홈, 의료 기기 등 일상생활 곳곳에서 활용되고 있다. 또한 빅데이터의 활용과 응용은 사회·경제·정치·법률 등 다양한 분야에서 영향을 미치고 있다. 이미 주요 선진국에서는 자율주행 및 얼굴 인증 시스템의 실용화에 따른 안전성과 개인정보 보호를 위해 통신·산업·의료·금융 분야 등에서 데이터 보호와 프라이버시 보호를 위한 규제와 인공지능에 의한 차별의 위험성, 인공지능 시스템에 대한 책임의 문제, 인공지능의 윤리 및 인공지능 정책 시스템 구축을 위한 현행 법제도의 조정을 위한 논의가 지속되고 있다. 한편 독일은 연방정부와 주 정부, 연구기관, 산업, 대학 등과 연계하여 고도의 인공지능 연구 인프라(Infra)를 구축하고 인공지능 기술이 가져올 이익과 편리함을 유용하게 활용하기 위하여 사회·경제·산업 분야 등에서 폭넓게 전개해 나가고 있다. 한편으로 인공지능 기술의 대두로 사회·윤리·법제도·문화 등 다양한 분야에서 초래될 변화에 대응하고 인공지능이 초래할 위험성에 대비하고자 법제도의 검토를 진행하고 있다. 우선 기술적 표준화를 통한 인공지능의 설계·제조·활용 등 일련의 과정에서 필요한 윤리와, 사회·경제·문화 분야 등 학제적 대화를 바탕으로 인공지능의 사회적 영향, 기술에 의한 미래의 영향에 관한 융합의 확대, 인공지능 기술이 인간의 안전을 위협하지 않도록 인공지능을 설계하는 논의를 지속하고 있다. 본고에서는 인공지능의 발전과 함께 기존의 사회적 가치관, 법제도와 충돌하지 않고 인공지능 기술을 활용할 수 있는 법제도의 정착을 위하여 독일의 인공지능 정책을 참조하여 부족하지만 우리나라의 인공지능 정책의 방향을 제시하는 논의의 장을 마련하고자 하였다.