인공지능에 의한 계약체결과 계약법이론 - 인공지능을 활용한 자율적 시스템에 대한 대리법의 적용 가능성을 중심으로 -

Considering that various incidents or problems are occurring in relation to artificial intelligence at present, it is natural to ask what is the best civil law system to attribute a third party's expression to the operator. Accordingly, the opinion of admitting artificial intelligence as an independent actor as a subject of expression that can attribute rights or direct or analogical application of proxy law has been argued from the initial discussion on autonomous systems using artificial intelligence. There is no doubt that this view is the most appropriate system among civil law systems, but various questions are being raised. While it may be reasonable to impose liability or liability for illegal activities caused by an AI-enabled system, the purpose of this paper is to impose appropriate and reasonable restrictions on contract responsibility for autonomous systems in the context of contract law. Unlike illegal activities, if the other party recognizes that autonomous systems are being used in the area of contract law, it should pay attention to them and sign contracts using autonomous systems, and in this respect, which requires rational adjustment of interests between the operator and the other party. Consequently, the ultimate purpose of this paper is to review the contract law scheme through the application of the proxy law, which may limit the operator's contract respon- sibility if the autonomous system enters into a contract differently than the operator's expression. First of all, in order for the proxy law to be applied, it is necessary to disclose to the other party that an autonomous system is being used as an agent. Otherwise, only the law on representative can be applied. Furthermore, since rights capabilities are not currently recognized in autonomous systems, it is difficult to recognize the system as the subject or agent of the expression of intent. Therefore, direct application of proxy laws is difficult to recognize. In addition, an autonomous system receives in principle comprehensive delegation from its operator, such as a blank, but technically, the system can implicitly limit the scope of agency rights, such as limiting money. Thus, the contract signed by an autonomous system is, in principle, comprehensive authorization, so it is likely that in principle, a representation proxy will be established. However, a representation proxy is not established if there is no justification for the other party to trust that the autonomous system has entered into the contract within the scope of the agency. Consequently, the analogical application of the proxy law provides the most rational model to attribute the contract signing action performed by the autonomous system to the responsibility of the autonomous system operator. However, if the autonomous system takes responsibility as an unauthorized agent, the problem remains that the autonomous system has no responsibility, such as forming a separate liability or insurance policy, so more discussions will be needed to overcome these limitations.