Earticle

이글은 인공지능과 관련된 규범의 형성에 관해서 살펴보고자 한다. 그동안 인공지능과 관련된 논의는 주로 현행 법규범에 따라 인공지능을 규범적으로 특히 법규범으로 평가를 한 연구성과가 많았다. 인공지능을 어떻게 활용하고자 하는 것인가에 대한 논의가 집중되어 있었다고 생각된다. 하지만 인공지능 기술이 발전함에 따라 인공지능의 활동 영역은 점점 더 확산될 것이므로, 이러한 변화에 발맞추어 새로운 규범 형성이 예상된다. 이글에서는 인공지능과 관련하여 윤리규범과 법규범의 형성에 관해서 살펴보지만, 주로 윤리규범을 중심으로 살펴보고자 한다. 법규범은 유럽연합에서 그동안 진행되었던 입법지침 등과 우리나라의 로봇법과 관련된 법률을 간략히 소개하였다. 윤리규범은 로봇윤리헌장, 섹스로봇 등을 중심으로 그 내용과 의견을 살펴보았다. 로봇윤리헌장은 로봇기술과 관련된 기술규범을 중심으로 한 것이다. 섹스로봇은 섹스로봇이 음란물인지 또는 기존의 성규범, 사회적 인식과 충돌이 발생하는지 여부에 관해서 간략히 소개하였다. 아직도 인공지능의 등장에 따라 윤리규범이나 법규범과 관련된 논의는 충분하다고 말할 수 없다. 인공지능과 관련된 윤리규범이라고 하였지만, 사실은 인간에 대한 규범, 인간행위에 대한 규범이라고 말할 수 있다. 인공지능을 제작, 이용할 때에 인간이 지켜야 할 행위의무를 규정하고 있다는 것이 유의할 점이다. 사회환경의 변화에 따라 법규범이나 윤리규범의 변화가 항상 있었다. 마찬가지로 인공지능의 등장에 따라 인간규범에 변화가 발생하거나 또는 인공지능 규범의 등장도 고려할 수 있다. 인공지능이 단순한 권리객체 또는 대상으로서 아니라 법률생활의 한 부분을 담당하는 주체로 등장하는 것이다.

This article will examine the formation of norms related to artificial intelligence. In the meantime, there have been many research results that conducted normative evaluation in accordance with the current legal norms in relation to artificial intelligence. As AI technology develops, the field of AI activity will spread more and more, so it is expected that norms will be formed in line with these changes. In this article, I looked at the formation of ethical standards and legal norms, but we will mainly focus on ethical standards. In relation to legal norms, I introduced the legislative guidelines that have been carried out in the European Union. In this article, regarding the Code of Ethics, the contents and opinions were reviewed focusing on the Robot Ethics Charter and Sex Robots. The Robot Ethics Charter is centered on the technical norms related to robot technology. Sex robots briefly introduced whether sex robots are pornography or whether they conflict with existing sexual norms and social perceptions. It can be said that code of ethics related to artificial intelligence, but in fact, is a norm for human being. It is important to note that when manufacturing and using artificial intelligence, it stipulates the duty of action to be observed by human being. In accordance with changes in the social environment, there have always been changes in laws and ethical standards. Similarly, with the advent of artificial intelligence, human norms may be changed, or the emergence of artificial intelligence norms may be considered. Artificial intelligence is emerging as a subject in charge of a part of legal life, not as a mere right object.

본 연구에서는 새로운 교사도우미 로봇에 초점을 맞추어 클라우드 기반의 교육서비스 모델을 연구하여 서 버영역에 적용하고, 클라이언트 영역에서는 교사도우미 로봇을 초등학교 교실환경에 적용하여 어학교육 서비스 플랫 폼으로 활용하고자 한다. 새로운 사물인터넷(IoT)기술 접목을 통해 쾌적한 스마트 교실환경을 만들고 다양한 미디어 에 대한 인터페이스를 지원하도록 한다. 이러한 목적의 달성을 위해 광범위한 선행연구와 사례분석을 통해서 서비스 모델구축에 필요한 기본적인 요구조건을 정리하였다. 임베디드 기반의 영상인식, 음성인식, 자율주행은 물론 디스플 레이, 터치스크린, IR센서, GPS, 온습도 센서에 대한 기술을 광범위하게 적용하여 서비스를 완성하도록 한다. 본 연 구결과의 가장 핵심적인 시사점은 클라우드 기술을 활용한 최적화된 플랫폼에 로봇러닝 및 IoT, BIM기술 융합을 통 한 지능형로봇기반의 스마트 교실구축 가능성 제시에 있다고 본다.

This study focuses on the new teaching assistance robot platform and the cloud-based education service model to support the server. In the client area we would like to use the teacher assistant robot in elementary school classrooms to utilize the language education service platform. Emerging IoT technology will be adopted to provide a comfortable classroom environment and various media interfaces. Extensive precedent review and case study have been conducted to identify basic requirements of proposed service platform. Embedded system and technology for image recognition, speech recognition, autonomous movement, display, touch screen, IR sensor, GPS, and temperature-humidity sensor were extensively investigated to complete the service. Key findings of this paper are optimized service platform with cloud server system and possibilities of potential smart classroom with intelligent robot by adopting IoT and BIM technology.

본 논문은 로봇의 도덕적 지위로서 권리 부여에 대한 비판적 검토 및 대안을 모색한다. 일반적으로 로봇의 도덕적 지위를 논할 때 주로 활용되는 서구의 ‘권리’ 개념은 권리를 가진 자들 간의 도덕적 이해관계 대립으로 인해 인간과 로봇 사이에 잠재적 갈등을 일으킬 수 있 다. 따라서 권리를 부여하는 것만이 로봇의 도덕적 지위를 다루는 유일하고도 적절한 방법이 아님을 주장하고 새로운 대안적 개념을 제시하고자 한다. 그 대안으로 로봇의 도덕적 지위는 '권리자(權利者, rights-bearers)'가 아니라 유학(儒學)에 기반한 '예자(禮者, rites-bearers)'가 되어야 함을 논증할 것이다. 먼저 ‘예(禮)’와 ‘예자(禮者)’의 개념을 논어를 통해 이해하고, 이를 서구의 ‘권리’와 ‘권리자’ 개념과 대조한다. 다음으로 ‘예자’로서 로봇의 지위에 대한 유 가적 관점 및 이의 정당성과 현실적인 함축을 논한다. 결론적으로 로봇의 도덕적 지위를 ‘예 자’로 간주하는 것은 인간과 로봇 그리고 공동체의 존재목적을 고려한 것으로 이는 공동체 내에서 인간과 로봇의 역할-의무를 다하게 하여 궁극적 상호작용으로서 조화와 협력을 이끌 수 있을 것이다.

This study provides alternatives through a critical review of the granting of rights as a moral status to robots. The concept of 'rights', which is mainly used in discussions on the moral status of robots, shows that granting rights is not the appropriate way to deal with the moral status of robots, since there are concerns about potential conflicts between humans and robots due to conflicting interests of rights holders. As an alternative, based on Confucian studies, we argue that the concept of moral status which can be applied to robots is ‘rites-bearers,’ not ‘rights-bearers.’ First, we define the concepts of ‘rite’ and ‘rites-bearers’ through the Analects of Confucius and contrast them with the Western definitions of ‘rights’ and ‘rights-bearers.’ Next, we discuss the Confucian perspective on the status of robots as ‘literates’ and its legitimacy and realistic implications. In conclusion, considering the moral status of robots as ‘yeja’ takes into account the purpose of the existence of humans, robots, and the community, and this can lead to harmony and cooperation as the ultimate interaction by having humans and robots fulfill their roles and duties within the community.

여성 사용자의 지능형 홈 제품 구매 성향을 분석하 고, 기능 디자인, 외관 디자인, 사용자 경험 디자인 등 다양한 측면에서 디자인 전략을 제안한다. 이를 통해 향후 제품 개발에 참고할 수 있는 전략을 제공하고, 시 장 경쟁력 강화를 목표로 한다. 본 연구는 SICAS-AHP 결합 모델을 기반으로 사용자 분석을 진행하였다. SICAS 모델을 활용해 사용자의 인식, 관심, 소통, 행 동, 공유 단계를 분석하고, 이를 AHP 계층 분석법과 결합하여 선호도 가중치를 산출하였다. 이 데이터를 바 탕으로 설계 제안을 도출하였다. 관찰법을 통해 여성 사용자의 집안 청소 행동과 제품 사용 습관을 조사하여 잠재적 수요와 문제점을 도출하였다. 또한 다양한 연령 대와 직업군의 여성 사용자들과 심층 인터뷰를 진행해 스마트 청소 로봇에 대한 요구, 기대, 구매 고려 요소 를 파악하였다. 설문지 작성 및 데이터 수집을 위해 관 찰 및 인터뷰 결과를 기반으로 설문 설계를 진행하였으 며, SICAS 모델의 각 단계(인식, 관심, 소통, 행동, 공 유)를 중심으로 데이터를 수집하고 이를 분석 및 요약 하였다. 조사 대상은 지능형 로봇 청소기이며, 여성 사 용자가 제품 구매 시 중요하게 고려하는 선호 요소를 분석하였다. SICAS 모델의 모든 요소를 계량화한 뒤, AHP 계층 분석법을 활용해 요소 모델을 구축하고, 선 호 가중치를 평가하여 핵심 요소를 도출하였다. 도출된 핵심 요소를 기반으로 구체적인 디자인 전략을 제안하 였다. 마지막으로, SICAS-AHP 결합 모델을 활용해 지 능형 청소 로봇의 디자인 실습을 완료했으며, 유사 제 품 개발에 참고할 수 있는 자료를 제공하였다.

This study mainly aims to understand the purchasing preferences of female users for intelligent floor cleaning robots, and propose design strategies including product function design, appearance design, user experience design, and other suggestions to improve the plan, providing design strategies for the development of related products in the future and enhancing market competitiveness. Use the SICAS-AHP combined model for user analysis, analyze users' perceptions, interests, communication, actions, and sharing through the SICAS model, and then combine AHP (Analytic Hierarchy Process) to analyze p reference weights and develop design recommendations. By using observation methods to observe female users' cleaning behavior and product usage habits at home, potential needs and pain points can be identified. Conduct in-depth inter views with female users of different age groups and professions through interview methods to u nderstand their needs, expectations, and purcha sing considerations for intelligent floor cleaning robots. Design a survey questionnaire and colle ct data based on SICAS. Analyze and summarize the collected research results from five aspects of the model. The survey focuses on intelligent floor cleaning robots and analyzes and summarizes the preference factors of female users when purchasing them. Quantify the various factors in the SICAS model, and then construct a factor model through AHP tomographic analysis. Conduct preference weight analysis through evaluation, select important factors based on weight ranking, and propose corresponding design strategies. Finally, an intelligent sweeping robot design practice was completed through the combination of SICAS-AHP model, providing reference f or the development of similar products.

기존 용접로봇제어기는 제조사별 프로토콜이 상이하며, 전용소프트웨어에 의해서만 제어와 모니터링이 가능하고, 판매가격 또한 고가로 도입 확대에 걸림돌이 되고 있으며, 기존 제품은 용접 시공 환경에 따른 전류, 전압과 압력 등에 대한 수동 조작방법으로 조작자의 숙련도에 의지하고 있는 실정이다. 그리고 기 설치된 생산관리시스템과 연동 이 되지 않아 불량률 증가와 관리인력 및 자원의 손실이 매우 크다. 따라서 본 논문은 이러한 문제점을 해소하고자 용접 공정 제어시스템을 빅데이터 기반의 스마트 팩토리화 하고, 지능형 센싱 에이전트, 생산관리시스템 연동 API, 자동제어장치가 산업기술용 네트워크로 연결된 새로운 기술을 제안하고 실증 수준의 시험평가를 실시하여 그 효용성을 입증하였다.

Existing welding robot controllers have different protocols for each manufacturer, can be controlled and monitored only by dedicated software, and the selling price is also expensive, which is an obstacle to the expansion of introduction, and existing products are current, voltage, and pressure depending on the welding construction environment. It is a situation that relies on the skill level of the operator as a manual operation method. And because it is not linked with the already installed production management system, the defect rate increases and the loss of management manpower and resources is very large. Therefore, in order to solve these problems, this paper turns the welding process control system into a smart factory based on big data, and proposes a new technology in which intelligent sensing agents, production management system interworking APIs, and automatic control devices are connected through a network for industrial technology. And proved its effectiveness by conducting empirical level test evaluation.

R-Learning is en educational system that makes use of robot functions, contents and IT convergence. R-Learning system was used in 1,843 kindergartens(2.147 EA) as of October, 2013. To accelerate the universalization of R-Learning, the related laws and system should be reviewed to seek workable reform measures. The purpose of this study was to examine an act on the development of the e-learning industry and the promotion of e-learning and an intelligent robot act in an effort to offer an overview of the legal system related to R-Learning. Besides, the regulations of the copyright law and the personal information protection law that were associated with R-Learning were reviewed to expedite the vitalization of R-Learning.

Coverage algorithm is one of the core technologies required for intelligent robots such as cleaning robots, harvesting robots, painting robots and lawn mowing robots. Although many smart coverage algorithms have been proposed, to the best of our knowledge, all of them make the same assumption that they have sufficient time to cover the entire target area. However, the time to completely cover the whole target area may not always be available. Therefore, in this paper, we propose another new coverage scheme, which we call the DmaxCoverage algorithm that decides the coverage path by considering the deadline for coverage. This approach can be beneficial when the time for the coverage is not sufficient to cover the entire target area. Experimental results show that the DmaxCoverage algorithm outperforms previous algorithms for these situations.

This paper presents a framework for smart services of robots in an intelligent environment. Within such intelligent environment, the target platform consists of service robots and a framework for smart services including task allocation and task scheduling. For task allocation, we use an auction-based method and a knapsack problem algorithm. In this paper, the characteristics of the algorithm for allocation balancing are to delegate executing task to another robot, to reallocate the delayed task to more than one robot, and to withdraw overallocated robots from the current task.

Intelligent robots are being introduced and utilized to create new values in diverse fields such as medical care, national defense, education, and manufacturing. However, while the technical aspects of intelligent robots are spreading, the business side is lacking in recognition. The purpose of this study is to identify the factors affecting product innovation through the introduction of intelligent robots in enterprises. Among the six variables secured from the technology, organization, and environment framework, only perceived direct usefulness and competitive pressure have significant effects on product innovation. The results imply that the effects of introducing robots are still in their early stage from a business perspective.

The intelligent robot industry is a complex which encompasses all fields of science and technology, and its marketability and industrial impact are remarkable. Major countries in the world have been strengthening their policies to foster the intelligent robot industry, but discussions on liability issues and legal actions that are accompanied by the related big or small accidents are still insufficient. In this study, therefore, the patent law by artificial intelligence robots and the legislation for relevant legal actions at the criminal law level are presented. Patent law legislation by artificial intelligence robots should comply with the followings. First, the electronic human being other than humans ought to be given legal personality, which is the subject of patent infringement. Even if artificial intelligence has legal personality, legal responsibility will be varied depending on the judgment of whether the accident has occurred due to the malfunction of the artificial intelligence itself or due to the human intervention with malicious intention. Second, artificial intelligence as a subject of actors and responsibility should be distinguished strictly; in other words, the injunction is the responsibility of the intelligent robot itself, but the financial repayment is the responsibility of the owner. In the criminal law legislation, regulations for legal punishment of intelligent robot manufacturing companies and manufacturers should be prepared promptly in case of legal violation, by amending the scope of application of Article 47 (Penal Provisions) of the Intelligent Robots Development and Distribution Promotion Act. In this way, joint penal provisions, which can clearly distinguish the responsibilities of the related parties, should be established to contribute to the development of the fourth industrial revolution.

For action selection as well as learning, simple associations between stimulus and response have been employed in most of literatures. But, for a successful task accomplishment, it is required that an animat can learn and express behavioral sequences. In this paper, we propose a novel action-selection-mechanism to deal with sequential behaviors. For this, we define behavioral motivation as a primitive node for action selection, and then hierarchically construct a network with behavioral motivations. The vertical path of the network represents behavioral sequences. Here, such a tree for our proposed ASM can be newly generated and/or updated, whenever a new sequential behaviors is learned. To show the validity of our proposed ASM, three 2-D grid world simulations will be illustrated.

We've been following research on the obstacle avoidance that is based on fuzzy control. We previously proposed a new method of automatically generating membership functions, which play an important role in improving accuracy of fuzzy control, by using genetic programming (GP). In this paper, we made two improvements to our proposed method, for the purpose of achieving better intelligence in fuzzy robots. First, the mutation rate is made to change dynamically, according to the coupled chaotic system. Secondly, the population partitioning using deme is introduced by parallel processing. The effectiveness of these improvements is demonstrated through several computer simulations.

This paper proposes an intelligent sliding mode control method for robotic systems with the unknown bound of model uncertainties. In our control structure, the unknown bound of model uncertainties is used as the gain of the sliding controller. Then, we employ the function approximation technique to estimate the unknown nonlinear function including the width of boundary layer and the uncertainty bound of robotic systems. The adaptation laws for all parameters of the self-recurrent wavelet neural network and those for the reconstruction error compensator are derived from the Lyapunov stability theorem, which are used for an on-line control of robotic systems with model uncertainties and external disturbances. Accordingly, the proposed method can not only overcome the chattering phenomenon in the control effort but also have the robustness regardless of model uncertainties and external disturbances. Finally, simulation results for the five-link biped robot are included to illustrate the effectiveness of the proposed method.

This paper proposes a method for the intelligent load distribution of two cooperating robots(TCRs) using fuzzy logic. The proposed scheme requires the knowledge of the robots' dynamics, which in turn depend upon the characteristics of large flat panel displays(LFPDs) carried by the TCRs. However, the dynamic properties of the LFPD are not known exactly, so that the dynamics of the robots, and hence the required Joint torque, must be calculated for nominal set of the LFPD characteristics. The force of the TCRs is an important factor in carrying the LFPD. It is divided into external force and internal force. In general, the effects of the internal force of the TCRs are not considered in performing the load distribution in terms of optimal time, but they are essential in optimal trajectory planning; if they are not taken into consideration, the optimal scheme is no longer fitting. To alleviate this deficiency, we present an algorithm for finding the internal-force (actors for the TCRs in terms of optimal time. The effectiveness of the proposed system is demonstrated by computer simulations using two three-joint planner robot manipulators.

Artificial intelligence (AI) education companion robots are becoming vital tools for personalized, interactive learning by integrating AI, robotics, and educational content, yet the interplay of perceived intelligence and warmth (key dimensions from Mind Perception Theory) and their impact on user satisfaction in this unique educational context remain insufficiently understood. This study aimed to empirically test the hypothesis that simultaneously enhancing intelligent and warm characteristics maximizes user experience in educational AI, where balancing cognitive support and social interaction is crucial, by investigating the interaction between these perceptions and their influence on user satisfaction. Through usability testing and questionnaires with 352 users of the WhalesBot AI education companion robot, user perceptions of intelligence, warmth, empathy, value, and their effects on satisfaction were analyzed. Findings revealed that perceived intelligence positively influences user satisfaction, partially mediated by perceived empathy and value. However, a “warm-intelligent paradox” emerged: perceived warmth significantly moderated the intelligence-empathy link but not the intelligence-perceived value link, indicating warmth does not uniformly enhance intelligence's positive effects on value. This challenges the “more humanized is better” design concept. These results highlight the necessity for context-sensitive design strategies, suggesting user satisfaction in AI educational tools is optimized by carefully balancing perceived intelligence and warmth based on interaction goals, rather than their indiscriminate maximization.

This paper presents a robust control method via the intelligent adaptive backstepping design technique for stable walking of nine-link biped robots with unknown model uncertainties and external disturbances. In our control structure, the self recurrent wavelet neural network(SRWNN) which has the information storage ability is used to observe the uncertainties of the biped robots. The adaptation laws for all weights of the SRWNN are induced from the Lyapunov stability theorem, which are used for on-line controlling biped robots. Also, we prove that all signals in the closed-loop adaptive system are uniformly ultimately bounded. Through computer simulations of a nine-link biped robot with model uncertainties and external disturbances, we illustrate the effectiveness of the proposed control system.

지능 로봇의 자율성은 인공지능의 발달정도에 의존할 수밖에 없다. 특히, 상당한 수준의 자율성을 갖춘 지능 로봇의 작동은 그 개발자 또는 발명자의 지배를 벗어난 것으로 평가될 여지가 있으며, 이러한 사정으로 인하여 로봇의 부적절한 작동결과가 타인에게 손해를 발생시킬 경우, 그에 대한 손해배상책임을 누구에게 부담시킬 것인지가 해결해야 할 과제로 대두될 수 있다. 이에 기존의 실정법이 이와 같은 인간과 로봇 사이의 상호작용을 규율하기에 충분한 것인지 아니면 로봇에게 법인격 내지 권리를 인정할 필요가 있는 것인지가 문제로 된다. 다만, 현행법상으로는 물건으로 다루어지는 단순 로봇에 대해 인공지능이 추가되는 경우 현재의 실정법이 그와 같은 인공지능형 로봇을 규율하기에 적합한 것으로 평가된다. 즉, 지능형 로봇이라고 하더라도 인간에 의한 통제가 가능한 수준이라면 현재의 책임법이 그대로 적용될 수 있다는 것이다. 그러나 지능형 로봇이 자율성을 가진 존재로 진화할 경우에는 특수한 책임모형을 모색할 필요가 있을 것이다. 현행 프랑스법에서의 규정태도를 종합적으로 분석한다면, 현재의 기술수준에서의 지능 로봇의 작동으로 인한 손해발생 상황에서의 규율에 있어 충분히 대처할 수 있으며 이러한 의미에서는 ‘법적 흠결’이 크지 않은 것으로 평가할 수도 있다. 다만, 강인공지능을 탑재한 자율적 지능 로봇의 출현은 아직까지 시간을 두고 지켜보아야 할 상황이라고는 하나, 그와 같은 자율적 지능 로봇으로 향하는 인공지능 기술의 개발이 나날이 향상됨으로 인하여 단순히 자동화된 로봇 이상의 자율성을 지닌 지능 로봇이 시장에 출시되고 있는 상황이다. 따라서 그와 같은 수준의 지능 로봇의 작동으로 인한 악결과를 규율하기 위해서는 현재의 각국의 법령들에서도 다소의 흠결이 발견되는 상황이라고 할 것이다. 이러한 의미에서 2020년 10월의 인공지능 운영자의 책임에 관한 유럽의회 결의안의 태도는 향후 자율적 지능 로봇의 출현을 준비하는 과정에서의 책임모델로 삼을 수도 있을 것이나, 동 결의안 또한 인공지능 운영자만을 수범자로 하고 있다는 점에서 한계가 있으며, 그 내용면에서는 프랑스에서의 Badinter법과 유사하다고 할 수 있다. 이상과 같은 점들을 종합적으로 고려할 때, 프랑스에서의 동물점유자의 책임 및 Badinter법이 향후 자율적 지능 로봇의 출현을 대비한 책임모델로서 보다 적합한 것으로 평가할 수 있을 것이다.

The autonomy of intelligent robots generally depends on the degree of development of artificial intelligence. In particular, the operation of an intelligent robot with a significant level of autonomy may be evaluated as outside the control of its developer or inventor, and if the improper operation result of the robot causes damage to others due to such circumstances, it can be difficult to determine who should be held responsible. Therefore, the question arises as to whether the existing positive law is sufficient to regulate such interactions between humans and robots, or whether it is necessary to recognize legal personality or rights to robots. However, since robots are treated as objects under the current law, even if artificial intelligence is added to the robot, it is evaluated that the current law can regulate the situation caused by the malfunction of the artificial intelligence robot. In other words, the current civil liability law can be applied as long as an intelligent robot can be controlled by humans. However, if intelligent robots evolve into beings with autonomy, it will be necessary to establish a special model of responsibility. If we comprehensively analyze the regulatory attitudes in the current French law, it can sufficiently regulate the situation of damage caused by the operation of the intelligent robot at the current technological level, and in this sense, it can be evaluated that the ‘legal flaw’ is not large. Although the emergence of autonomous intelligent robots equipped with strong artificial intelligence is still a situation that needs to be watched over time, intelligent robots with greater autonomy than simple automated robots are coming to market, as artificial intelligence technology develops day by day. Therefore, it can be said that there are some flaws in the current laws and regulations of each country in order to regulate the bad consequences of the operation of these intelligent robots. In this sense, the attitude of the European Parliament resolution on the responsibility of AI operators in October 2020 can be used as a model of responsibility in the process of preparing for the advent of autonomous intelligent robots in the future, but this resolution has limitations in that it only targets AI operators. Considering the above points collectively, it can be evaluated that the liability of animal occupants and the Badinter Act in French law are more suitable as a model of responsibility for the emergence of autonomous intelligent robots in the future.

이 연구는 어린이집 만 5세반 교실에 인공지능 로봇을 투입하여 자연스러운 환경에서 유아들이 로봇과 어떻게 상호작용하는지, 그리고 시간의 흐름에 따라 상호작용 양상이 달라지는지 살펴보는 데 목적이 있다. 이에 인공지능 로봇 알파미니(ALPHA MINI)를 만 5세반 어린이집 2학급(유아 33명, 교사 3명)에 투입하여 3개월간 유아, 교사, 로봇간의 상호작용을 관찰하였으며, 로봇 활용 후 교사 면담을 실시하였다. 관찰기록과 면담자료는 주제분석 방법을 적용하여 분석하였고, 연구 문제에 관련된 결과를 도출하였다. 주제분석 결과, 교실 내에서 유아는 로봇과 주로 단순한 언어를 사용하여 오락 위주의 상호작용, 정보습득용 상호작용, 정서적 상호작용을 하는 것으로 나타났다. 또한 시간의 흐름에 따른 유아와 로봇의 상호작용 양상을 살펴보았을 때, 유아와 로봇 간의 상호작용 양상이 눈에 띄게 달라지지는 않았다. 하지만 로봇과의 상호작용을 주도하는 유아가 순환하는 것으로 나타났고, 로봇을 매개로 유아와 유아, 유아와 교사 간의 상호작용 양상이 변화하는 것으로 나타났다. 이를 바탕으로 유아반 교실에서 인공지능 로봇을 활용하기 위해 선행되어야 할 기술적 지원, 교사교육, 인공지능 윤리에 대한 논의를 제시하였다.

The purpose of this study is to introduce artificial intelligence (AI) robots to daycare centers to examine how children interact with robots in a natural environment and whether these interaction patterns change over time. Accordingly, the AI robot ALPHA MINI was introduced into two classes (33 children and 3 teachers) at daycare centers for 5 years olds to observe interactions among children, teachers, and a robot for three months. In addition, interviews with teachers were conducted after using the robot. The findings of the study revealed that children mainly use simple language with robots to engage in entertainment-oriented interactions, information acquisition interactions, and emotional interactions. Furthermore, while the interaction between children and robots remained relatively stable over time, it was noted that children leading the interaction with robots circulated, and the interaction patterns between peers mediated by robots has changed. Based on these findings, we highlight the need for prior deliberations on technical support, teacher education, and AI ethics as prerequisites for the integration of AI robots into early childhood classrooms.

In this study, we analyzed technological research trends in intelligent robots and identified key technologies using keyword network analysis. Frequency and centrality analyses were performed on 320 papers published in Korea since 2000. Related research has increased since 2010, and active research has been conducted especially since 2020. Recognition technology, including sensor, SLAM, and object recognition, accounted for the highest proportion, followed by intelligence technology, integration and interaction technology, and motor technology. High-frequency keywords such as mobile robot, AI, autonomous driving, reinforcement learning, and SLAM also showed high network centrality. This study suggests that future research should focus on key issues such as safety, reliability, human-robot interaction, and standardization.

在人工智能技术迅猛发展的背景下，智能机器人的风险监管和治理问题已经上升为一个全球性议题。基于科学技术与社会（STS）理论，协同治理模式已经成为智能机器人风险治理领域内的一个共识性理念。然而，关于如何在实施途径、内在机制和治理实践上进行有效监管，目前还缺乏系统性的理论支持和实践指导。现有的监管部门往往各自为政、孤立运作，未能有效实现跨领域的协调与整合。在实践改革中，普遍采用“分类”方法来应对复杂性和融合性的挑战。本文案例研究聚焦于智能机器人在中国的工业制造、医疗服务、仓储服务和餐饮服务四个领域的实际应用。通过对智能机器人在技术层面和业态层面的深入分析，并借鉴各领域现有的国际治理经验，旨在创新性地构建一个智能机器人综合治理体系，从而为监管实践者提供分类思路，以实现“共同治理、共享资源、融合发展”的治理模式。

Amid the rapid advancements in artificial intelligence (AI) technologies, the regulation and governance of risks associated with intelligent robots have emerged as a global concern. Grounded in the Science, Technology, and Society (STS) theory, the collaborative governance model has gained consensus as a pivotal framework for addressing these risks. However, there remains a lack of systematic theoretical support and practical guidance on effective regulatory approaches, internal mechanisms, and governance practices. Existing regulatory bodies often operate in silos, failing to achieve cross-sectoral coordination and integration. In response to these challenges, ‘classification-based’ methods are widely adopted in practice to address the complexity and convergence of intelligent robot applications. This study focuses on case analyses of intelligent robot applications across four sectors in China: industrial manufacturing, healthcare services, warehousing services, and food services. By conducting an in-depth examination of the technological and business dimensions of intelligent robots and drawing upon international governance experiences in these domains, this paper seeks to innovatively construct an integrated governance system for intelligent robots. The proposed framework aims to provide regulatory practitioners with a classification-based approach to realize a governance model characterized by ‘collaborative governance, shared resources, and integrated development.’