인공지능의 형법적 통제에 관한 고찰 - 인공지능기본법을 중심으로 -

Although Artificial Intelligence(AI) brings utility to human civilization, there is a risk of it being abused as a useful tool for various crimes. The recent rapid development of generative AI has caused controversy in various crimes such as the production and distribution of fake news, fraud, and sexual crimes. Europe has already enacted a law to regulate such AI, and Korea has also enacted the AI Framework Act in 2025, presenting standards for management of the AI industry. However, the effectiveness of the new Act 2025 is still questionable as it merely suggests a future direction rather than allowing the state to proactively prevent and actively regulate the risks of AI. The reason for this may be that it is difficult to clearly establish the nature of AI and that there were concerns that the development of related industries would be hindered. The AI Framework Act must be contained the following three things: Firstly, designation of prohibited AI, secondly, mandatory prior screening and certification of high-risk AI, and lastly, sanctions for breaches of duty by business operators should be strengthened. Rather, the current legislative trend is to create new punishment laws for each case, such as the regulation of deepfake pornography. There is concern that this method will lead to the enactment of punishment laws that lack theoretical review. Therefore, a theoretical review is required for detailed classification of high-risk AI, reasonable standards for duty of care under Product Liability Law, and distribution of responsibility among involved entities such as developers and manufacturers.