Scholarlink

<p>This paper examines the ethical and systemic implications of artificial intelligence (AI) in financial markets by developing a unified framework that integrates technical risk dynamics with societal outcomes. We introduce the Systemic Expected Shortfall with Ethical Dimensions (<InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(SE{S}_{AI,Ethical}^{\alpha }\)</EquationSource> </InlineEquation>), which extends traditional risk measures by incorporating distributional consequences such as inequality, bias, and disproportionate losses across stakeholders. The model identifies key AI-specific amplification channels—including speed and connectivity, bias and inequity exposure, reward hacking, model homogeneity, and orchestration fragility—through which systemic risk is generated and propagated. Illustrative market episodes, including the GameStop trading frenzy, the Flash Crash, spoofing practices, and the London Whale incident, are used to show how these mechanisms may operate in practice. The analysis shows that AI enhances efficiency under normal conditions but can amplify fragility and ethical harm under stress. The paper concludes by outlining governance implications for aligning AI-driven financial innovation with fairness, accountability, and systemic stability.</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

Artificial intelligence, ethical amplification, and systemic risk in financial markets

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年6月30日

查看全文
  • Jingrui Li

摘要

This paper examines the ethical and systemic implications of artificial intelligence (AI) in financial markets by developing a unified framework that integrates technical risk dynamics with societal outcomes. We introduce the Systemic Expected Shortfall with Ethical Dimensions ( \(SE{S}_{AI,Ethical}^{\alpha }\) ), which extends traditional risk measures by incorporating distributional consequences such as inequality, bias, and disproportionate losses across stakeholders. The model identifies key AI-specific amplification channels—including speed and connectivity, bias and inequity exposure, reward hacking, model homogeneity, and orchestration fragility—through which systemic risk is generated and propagated. Illustrative market episodes, including the GameStop trading frenzy, the Flash Crash, spoofing practices, and the London Whale incident, are used to show how these mechanisms may operate in practice. The analysis shows that AI enhances efficiency under normal conditions but can amplify fragility and ethical harm under stress. The paper concludes by outlining governance implications for aligning AI-driven financial innovation with fairness, accountability, and systemic stability.