Scholarlink

<p>The rapid global deployment of artificial intelligence (AI) systems has exposed a profound governance asymmetry. Frameworks designed within Western liberal democratic traditions—most prominently the European Union AI Act and the General Data Protection Regulation (GDPR)—encode assumptions of individual autonomy, personal data ownership, and consent-driven data processing that are epistemologically misaligned with the relational and collective understandings of privacy prevalent across large portions of the Global South. This misalignment is not merely a regulatory gap; it is a structural condition through which AI technologies, developed and governed in the Global North, can extract, infer, and commodify the collective knowledge, behavioral patterns, and cultural heritage of communities with no meaningful voice in the governance architecture that ostensibly protects them—a dynamic scholars have termed ‘digital colonialism.’ This perspectives paper argues that addressing this structural condition requires a paradigm shift toward Context-Aware Governance: a multi-pillar approach grounded in local epistemologies, collective consent mechanisms, expanded definitions of algorithmic harm, group-level technical privacy safeguards, and participatory policy co-design. Drawing on Ubuntu philosophy, Indigenous Data Sovereignty principles, and the CARE principles framework, we delineate five pillars through which Context-Aware Governance can be operationalized: the philosophical re-evaluation of privacy’s foundations; the institutionalization of collective consent through Community Data Trusts and tiered consent models; the redefinition of algorithmic harm to encompass collective, epistemological, and resource-depletion injuries; the development of group-level differential privacy and federated learning architectures; and the establishment of participatory Community Impact Assessments. We argue that the transition from the FAIR principles to the CARE principles represents a necessary normative shift in how global AI governance conceptualizes the relationship between data, communities, and power, and that the governance of AI in communal societies constitutes a central test of whether the global AI project can genuinely serve all of humanity.</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

Beyond individual consent: toward Context-Aware Governance frameworks for artificial intelligence in the Global South

  • 期刊论文

  • 发表日期：

    2026年7月6日

查看全文
  • Shihshuan Fang,
  • Shenghan Chen

摘要

The rapid global deployment of artificial intelligence (AI) systems has exposed a profound governance asymmetry. Frameworks designed within Western liberal democratic traditions—most prominently the European Union AI Act and the General Data Protection Regulation (GDPR)—encode assumptions of individual autonomy, personal data ownership, and consent-driven data processing that are epistemologically misaligned with the relational and collective understandings of privacy prevalent across large portions of the Global South. This misalignment is not merely a regulatory gap; it is a structural condition through which AI technologies, developed and governed in the Global North, can extract, infer, and commodify the collective knowledge, behavioral patterns, and cultural heritage of communities with no meaningful voice in the governance architecture that ostensibly protects them—a dynamic scholars have termed ‘digital colonialism.’ This perspectives paper argues that addressing this structural condition requires a paradigm shift toward Context-Aware Governance: a multi-pillar approach grounded in local epistemologies, collective consent mechanisms, expanded definitions of algorithmic harm, group-level technical privacy safeguards, and participatory policy co-design. Drawing on Ubuntu philosophy, Indigenous Data Sovereignty principles, and the CARE principles framework, we delineate five pillars through which Context-Aware Governance can be operationalized: the philosophical re-evaluation of privacy’s foundations; the institutionalization of collective consent through Community Data Trusts and tiered consent models; the redefinition of algorithmic harm to encompass collective, epistemological, and resource-depletion injuries; the development of group-level differential privacy and federated learning architectures; and the establishment of participatory Community Impact Assessments. We argue that the transition from the FAIR principles to the CARE principles represents a necessary normative shift in how global AI governance conceptualizes the relationship between data, communities, and power, and that the governance of AI in communal societies constitutes a central test of whether the global AI project can genuinely serve all of humanity.