ARTIFICIAL INTELLIGENCE (AI) AS AN EXISTENTIAL THREAT

THE ROLE OF THE EUROPEAN UNION AND ITS AI ACT IN TRANSNATIONAL POLICYMAKING

Abstract

Apart from various benefits artificial intelligence (AI) brings for humanity, recent years have generated greater awareness of multifaceted dangers and threats it could represent. Among them, the most usually invoked are the disturbance of the labor market, potential threats for the human creativity and protection of intellectual property rights, human alienation, risks of manipulative media content and cross-border cyber interference, and, finally, the greatest threat of all: the existential one. Already well-known and thoroughly discussed potential causes of catastrophic outcomes for humanity include all-encompassing threats, such as climate change, nuclear war, or pandemics. While terminology varies, threats that could result in partial or complete destruction of humanity are typically termed “existential risk/threat” or a threat to “existential security”. On the list of possible threats, one of the most recent entries is the danger to humanity’s survival posed by the actions of some advanced, potentially self-aware, and highly destructive AI. The EU Regulation 2024/1689 of 13 June 2024 laying down harmonised rules on artificial intelligence (EU AI Act) is the first supranational legally binding act of this kind. However, due to various reasons – such as the EU’s internal political dynamics and the limitations imposed by its decision-making processes – the EU AI Act is the smallest common denominator of its Member States’ regulatory intentions, leading to controversial, complex and often unenforceable legal norms. While Art. 5 of the EU AI Act is dedicated to different prohibited AI practices, the EU legislation does not offer an applicable regulatory framework dedicated to existential threat from AI. At the same time, different trans-national initiatives – such as Center for AI Safety, Future of Life Institute, AI Red Lines – have already proposed different policy and regulatory measures to mitigate the risk of extinction from AI. The article first focuses on the notion of existential threat from AI and the potential self-awareness AI could achieve (Chapter 2), before turning to the issue of global regulatory initiatives aiming to mitigate the most significant AI risks (Chapter 3); then it examines the existing EU legislation, as well as the Union’s current and potential role in transnational AI-related policymaking (Chapter 4). The methodological approach is based on dynamic and holistic legal content analysis of the EU legislation and policy documents, while the literature review is combined with references to relevant jurisprudence and insights into legal reasoning of policy makers and other relevant stakeholders. Finally, whenever appropriate, the analysis is based on convergence of legal studies with interdisciplinary explorations of political science, psychology, and technology.