디지털 어시스턴트를 통한 계약 체결과 소비자보호 - ELI DACC 모델법 해제(解題)를 중심으로 -

The rapid advancement of artificial intelligence (AI) has enabled the automation of contract formation and performance through digital assistants, fundamentally reshaping the structure of consumer transactions. At the same time, this automated environment raises novel legal challenges, including algorithmic opacity, conflicts of interest, manipulation risks, and gaps in liability. This article analyzes the regulatory framework and consumer protection implications of the Guiding Principles and Model Rules on Digital Assistants for Consumer Contracts (ELI DACC), which the European Law Institute (ELI) adopted in 2025. The ELI DACC conceptualizes digital assistants not as legal agents but as tools of the user, thereby adopting a principle of attribution under which their acts are, in principle, imputed to the user. However, attribution is limited where the assistant’s conduct exceeds the consumer’s reasonable expectations, preventing the excessive allocation of risk to consumers. A distinctive feature of the model rules is their practical incorporation of Lawrence Lessig’s “Code is Law” thesis through ex ante design requirements. These include mandatory functionalities such as parameter-setting and modification, contract-prevention mechanisms including confirmation or objection models, and the right to deactivate the assistant. In addition, the framework strengthens transparency and fairness by requiring machine-readable information, disclosure of conflicts of interest, and safeguards against manipulation of digital assistants. Through conformity standards and a risk-allocation approach, liability is primarily placed on providers as the “cheapest cost avoiders,” thereby facilitating effective remedies for financial losses arising from automated transactions. This study argues that the ELI DACC represents a proactive regulatory model addressing the regulatory lag between rapid technological development and existing private law doctrines. By integrating technical design requirements with judicial remedies, the model offers a preventive regulatory paradigm that accommodates innovation while preserving meaningful consumer control. The principles of attribution and reasonable expectations articulated in the ELI DACC provide a normative benchmark for ensuring legal certainty in the emerging ecosystem of autonomous AI agents. The analysis further suggests that the model rules offer valuable guidance for modernizing Korean consumer and e-commerce law to reflect the distinctive features of AI-driven contracting.