Artificial Intelligence: Risks, Governance, and the Path from Privacy Paradox to Digital Inevitability

Artificial Intelligence (AI), particularly Generative AI (GenAI) and Large Language Models (LLMs), is reshaping digital services while raising ethical, legal, cybersecurity, privacy, and alignment challenges. This narrative review provides an interdisciplinary synthesis of these challenges, examining GenAI-related risks, the evolving European Union regulatory landscape, including the AI Act, cybersecurity threats to AI systems, and core principles for trustworthy AI. Beyond this synthesis, the paper’s central contribution is a reconceptualisation of the privacy paradox: where data collection is structurally entrenched, non-participation is costly, and viable lower-data alternatives are lacking, the gap between privacy attitudes and disclosure behaviour can reflect structural constraints on participation as well as individual decisions, a conditional account we term “digital inevitability” (DIGINEV). GenAI does not create this condition but can intensify it by inferring sensitive information from indirect signals. We develop this into a conceptual framework that proposes how constraints on participation may interact with user resignation and identifies possible responses through regulation, design, collective action, and viable alternatives. The paper further discusses AI alignment as a supporting consideration in the governance of GenAI systems. The findings offer an integrated perspective for future research, regulatory development, and responsible AI deployment, anchored in the DIGINEV framework.