Scholarlink

<p>Questions of AI sovereignty are conventionally framed in terms of state capacity and geopolitical positioning, yet this article argues that such framings are analytically insufficient and socially consequential in ways that existing policy frameworks have not adequately addressed. Through a critical interpretive review of the Tony Blair Institute for Global Change’s influential framework which defines AI sovereignty as "strategic agency" within conditions of irreversible interdependence, the article identifies three interconnected limitations: the framework underspecifies how infrastructural power structurally bounds strategic choice; it fails to account for coercion pathways that only become visible under disruption; and in post-colonial and capacity-constrained contexts, it risks functioning as a discourse that manages rather than contests hierarchical structures in the global AI political economy. Crucially, the article proposes that AI governance is an irreducibly intermestic problem, simultaneously and inseparably international and domestic in character, extending Manning’s (1977) framework to the AI domain as an original theoretical contribution. This framing reveals that geopolitical dependencies and domestic accountability deficits are mutually constitutive: what states can choose internationally shapes what citizens can contest domestically, and vice versa. In response, the article proposes power-aware governance as a more adequate conceptualisation of AI sovereignty, operationalised through short-horizon stress-testing and scenario-based analysis. It further argues that established international relations concepts including bandwagoning, hedging, and soft power require substantive theoretical extension, and that future conceptual development must be intermestic by design if it is to capture the triangulated dynamics between states, platform firms and supranational bodies that define the AI landscape. The argument has direct relevance for democratic legitimacy, public accountability, and social justice in AI-dependent public services.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

AI sovereignty without power? Reviewing strategic agency, infrastructural dependence and digital imperialism

  • 期刊论文

  • 发表日期：

    2026年6月16日

查看全文
  • Afzal Ashraf,
  • Vito Veneziano

摘要

Questions of AI sovereignty are conventionally framed in terms of state capacity and geopolitical positioning, yet this article argues that such framings are analytically insufficient and socially consequential in ways that existing policy frameworks have not adequately addressed. Through a critical interpretive review of the Tony Blair Institute for Global Change’s influential framework which defines AI sovereignty as "strategic agency" within conditions of irreversible interdependence, the article identifies three interconnected limitations: the framework underspecifies how infrastructural power structurally bounds strategic choice; it fails to account for coercion pathways that only become visible under disruption; and in post-colonial and capacity-constrained contexts, it risks functioning as a discourse that manages rather than contests hierarchical structures in the global AI political economy. Crucially, the article proposes that AI governance is an irreducibly intermestic problem, simultaneously and inseparably international and domestic in character, extending Manning’s (1977) framework to the AI domain as an original theoretical contribution. This framing reveals that geopolitical dependencies and domestic accountability deficits are mutually constitutive: what states can choose internationally shapes what citizens can contest domestically, and vice versa. In response, the article proposes power-aware governance as a more adequate conceptualisation of AI sovereignty, operationalised through short-horizon stress-testing and scenario-based analysis. It further argues that established international relations concepts including bandwagoning, hedging, and soft power require substantive theoretical extension, and that future conceptual development must be intermestic by design if it is to capture the triangulated dynamics between states, platform firms and supranational bodies that define the AI landscape. The argument has direct relevance for democratic legitimacy, public accountability, and social justice in AI-dependent public services.