Scholarlink

<p>Nick Bostrom’s simulation argument proposes a striking trilemma: either almost no civilisation reaches a posthuman stage, posthuman civilisations almost never run detailed “ancestor simulations”, or almost all observers like us live in computer simulations. This paper builds on that framework but shifts the focus from metaphysics to institutions. Instead of asking whether we are simulated, it asks a more concrete question in applied AI governance and the ethics and political economy of digital worlds: who would actually pay to run a Matrix-like simulation, why, and under what constraints would they keep paying?. By a “Matrix-like simulation” I mean a high-fidelity, open-ended environment in which very large numbers of agents have the kinds of cognitive capacities and life-trajectories that would ordinarily make us treat them as serious moral patients, which is a stricter threshold than ordinary agent-based models or narrow digital twins. Drawing on existing work on simulations, digital twins, large-scale modelling, and AI governance, the paper introduces the BRACE framework, Benefits, Resources, Accountability, Conscience, Existential risk, as a way of analysing “Net Matrix Value” for different kinds of potential sponsors, including commercial firms, states, scientific consortia, safety organisations, and ideological or artistic actors. It then formulates a simple Matrix-Funding Inequality that clarifies when large-scale simulations are economically, politically, and ethically sustainable. The central claims are that: (1) plausible sponsors face strong cost, risk, and normative pressures that favour narrow, goal-directed simulations over open-ended Matrix worlds, and (2) reasoning about who would fund simulations illuminates current debates on AI rights, high-risk AI regulation, and cross-border governance of advanced AI systems. The aim is not to argue that we are simulated, but to use funding incentives as a bridge between simulation theory, AI ethics, and the political economy of emerging digital worlds.</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

Who will pay for the Matrix? Simulation sponsors, AI governance, and the ethics and political economy of digital worlds

  • 期刊论文

  • 发表日期：

    2026年2月22日

查看全文
  • Stephen Topazi

摘要

Nick Bostrom’s simulation argument proposes a striking trilemma: either almost no civilisation reaches a posthuman stage, posthuman civilisations almost never run detailed “ancestor simulations”, or almost all observers like us live in computer simulations. This paper builds on that framework but shifts the focus from metaphysics to institutions. Instead of asking whether we are simulated, it asks a more concrete question in applied AI governance and the ethics and political economy of digital worlds: who would actually pay to run a Matrix-like simulation, why, and under what constraints would they keep paying?. By a “Matrix-like simulation” I mean a high-fidelity, open-ended environment in which very large numbers of agents have the kinds of cognitive capacities and life-trajectories that would ordinarily make us treat them as serious moral patients, which is a stricter threshold than ordinary agent-based models or narrow digital twins. Drawing on existing work on simulations, digital twins, large-scale modelling, and AI governance, the paper introduces the BRACE framework, Benefits, Resources, Accountability, Conscience, Existential risk, as a way of analysing “Net Matrix Value” for different kinds of potential sponsors, including commercial firms, states, scientific consortia, safety organisations, and ideological or artistic actors. It then formulates a simple Matrix-Funding Inequality that clarifies when large-scale simulations are economically, politically, and ethically sustainable. The central claims are that: (1) plausible sponsors face strong cost, risk, and normative pressures that favour narrow, goal-directed simulations over open-ended Matrix worlds, and (2) reasoning about who would fund simulations illuminates current debates on AI rights, high-risk AI regulation, and cross-border governance of advanced AI systems. The aim is not to argue that we are simulated, but to use funding incentives as a bridge between simulation theory, AI ethics, and the political economy of emerging digital worlds.