생성형 인공지능(AI) 규제원칙에 관한 연구

In November of last year, 8 years after AI AlphaGo created a sensation by beating 9th dan Lee Se-dol completely at badulk in 2015, ChatGPT, an interactive artificial intelligence, appeared, shocking the world once again. With the advent of ChatGPT, humans are starting to discuss AI regulations in earnest, believing that the dangers of robots and artificial intelligence, which have only been seen in movies, will become a reality. Prior to the release of generative AI, the prevailing opinion was that AI ethics should be strengthened. Afterwards, when generative AI is released and the prospect that it will be a game changer prevails, major countries such as the EU, the US, and Korea are starting legal discussions on generative AI. While the EU is in the process of legislating first, the US and Korea have also prepared legislative bills and are discussing them in the National Assembly. However, it is important to understand the characteristics of generative AI and set the direction of regulation before the full-scale introduction of legal regulations. In general, new technology regulation can derive four standards or principles: content or characteristics of technology, timing and region of regulation, risk and reward according to regulation, step-by-step and comprehensive approach, and guarantee of professional and democratic participation. Generative AI has tremendous potential for innovation in that it can generate original and creative content, not only text, but also images, music, and even computer code, using provided data. It is difficult to regulate because it is difficult to predict the development direction or impact of generative AI technology, but considering the risks, it is also difficult to choose not to regulate at all. In the direction of regulation, it is important to consider the principle of proportional regulation considering principle-based regulation, negative regulation, harmonization of domestic and international regulation, and risk-based regulation.