Systemic Risks and Governance of Advanced Artificial Intelligence: A Structured Literature Review

Counts differ because each database indexes a different set of publications. We treat OpenAlex as the canonical count; Google Scholar is not shown (no API, and crawling it violates its ToS).

Advanced artificial intelligence (AI) is advancing at a pace that exceeds the capacity of the ethical, legal, and institutional systems designed to govern it. Although scholarly attention to specific AI-related risks has grown substantially, no integrative framework has yet systematically mapped the relationships among technical misalignment, competitive market pressures, political economy, and multilevel governance responses. This structured literature review addresses two questions: (1) What are the principal systemic risks arising from advanced AI development? and (2) Which layered governance interventions are most appropriate for addressing each category of risk? A structured narrative review was conducted across the Scopus, Web of Science, and IEEE Xplore databases, supplemented by grey literature from established AI research institutions. Inclusion criteria required primary engagement with AI risk, alignment, governance, or sociotechnical impact. Sources published between 2014 and 2025 were synthesized through thematic analysis, resulting in the identification of six risk categories and a four-level governance framework. The review identifies six interlocking systemic risks: (1) displacement of human control through misaligned optimization; (2) large-scale automation of cognitive labor; (3) competitive incentive traps that drive unsafe acceleration; (4) AI-mediated manipulation of language-based institutions; (5) emergent misaligned agentic behavior; and (6) dangerous concentration of data, infrastructure, and epistemic power. These risks are not discrete; rather, they are structurally interdependent. This paper makes three original contributions: a synthesized six-risk taxonomy; a four-level governance model spanning the individual, developer, state, and international levels; and a critical appraisal of the evidentiary basis for each risk claim. The analysis suggests that dangerous outcomes are structurally plausible, though not inevitable, and that coordinated multilevel governance is the decisive variable. Keywords: artificial intelligence; AI alignment; AI governance; labor automation; sociotechnical risk; political economy of AI; technology ethics