자율주행자동차의 형법적 쟁점 및 대응방안

Abstract

Autonomous driving vehicles have rapidly developed through data-based machine learning and deep learning based on artificial neural networks. Recent deep learning algorithms allow objects to be recognized through physical movement, and it is expected that autonomous driving vehicles will be able to understand the actual road conditions more naturally. How should the criminal law regulate autonomous driving vehicles? First of all, Risk criminal law focuses on preemptive measures to future potential risks. However, it will eventually lead to the functionalization or policyization of criminal law. On the other hand, the allowed risk theory makes it possible to take some risks arising in modern industrial societies. In the case of autonomous driving vehicles, the problem is to what extent can our society allow risks from errors in autonomous driving vehicles’ software? Although the 「Act on Promotion and Support of Commercialization of autonomous driving vehicles」 is enacted and implemented in 2019, it focuses on the development of autonomous vehicle technology and the establishment of infrastructure, and there are no regulations on the risks. Therefore, it is necessary to establish the permissible range of autonomous driving vehicle risks to ensure the safety of the people and to establish a legal basis for the preliminary regulation and management of risks exceeding them. ‘Mean Distance Between Failures(MDBF)’ can be used as a quantified basis for determining the safety of autonomous driving vehicles. And I propose to form a public-private cooperation committee with various members to ensure rapid legislative and policy response, and to mandate prior management of manufacturers and regular software inspections by users.