The autonomy, self-learning capacity, and opacity of Artificial Intelligence (AI) have presented significant challenges for the legal field. Traditionally, the law regulates human decisions and incentivises humans and legal entities such as corporations or societies. However, AI does not fit neatly into this framework, leading to the emergence of scholarly discussions on affording legal personhood to AI. Based on the recent book on the regulation of liability for AI, the author examines the arguments for and against treating AI as a legal person and alternative avenue. It explores historical precedents, such as the treatment of slaves under Roman law, and draws parallels to modern company law, where legal personhood is not necessarily binary—entities need not possess all rights or none at all. The talk also discusses the European Parliament’s initial recognition of the idea to grant legal personhood to AI and its subsequent rejection. Through this lens, the talk provides a nuanced analysis of how legal systems might adapt to address the complexities posed by AI. Moreover, the presentation discusses Bosnia and Herzegovina and the countries of the Western Balkans and how they can respond to these challenges.