Scholarlink

<p>The divergence between embodied AI (AI) and human cognition arises from differences in substrate, architecture, and developmental trajectories. Human intelligence emerges from carbon-based biochemical and evolutionary processes, while embodied AI develops through non-carbon substrates and distinct control systems. This contrast raises key questions: will embodied AI follow a trajectory toward Artificial general intelligence (AGI) comparable to that of carbon-based beings, or will it diverge into novel forms of intelligence? And if divergence is inevitable, what social and ethical consequences follow? Embodied AI may increasingly replicate three core cognitive stages that we propose—perception (“I perceive, therefore I exist”), reflexivity (“I reflect on what I perceive”), and Theory of Mind (“I consider what others perceive”). These advances point toward the possibility of Artificial general intelligence (AGI), but also heighten the risk of anthropomorphizing such systems, mistaking opaque intentions for human-like cognition. To look into this, we propose the metaethics of substrate framework, which holds that the ethical challenges and responsibilities associated with intelligence cannot be reduced to abstract measures of functional performance; here, we assume that they must also account for the material substrate through which cognition emerges. In this regard, Martha Nussbaum’s capabilities approach reminds us that even biological intelligence is plural and diverse; by extension, divergence between biological-based and non-biological embodied AI may prove even greater. Recognizing these differences is vital in the current race to AGI, where market pressures often eclipse ethical reflection. Therefore, this work argues that bridging embodied AI research with frameworks such as metaethics of substrate and the capabilities approach is essential for clarifying how these systems should be situated in society, what forms of agency they may possess, and how their divergence from human cognition can be managed responsibly. es and initials are correct.</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

On substrate meta ethics as a framework for understanding human and AI divergence in embodied artificial intelligence

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年8月4日

查看全文
  • Jorge Luis Morton,
  • Bernardo Bolaños Guerra,
  • Daniel Romero

摘要

The divergence between embodied AI (AI) and human cognition arises from differences in substrate, architecture, and developmental trajectories. Human intelligence emerges from carbon-based biochemical and evolutionary processes, while embodied AI develops through non-carbon substrates and distinct control systems. This contrast raises key questions: will embodied AI follow a trajectory toward Artificial general intelligence (AGI) comparable to that of carbon-based beings, or will it diverge into novel forms of intelligence? And if divergence is inevitable, what social and ethical consequences follow? Embodied AI may increasingly replicate three core cognitive stages that we propose—perception (“I perceive, therefore I exist”), reflexivity (“I reflect on what I perceive”), and Theory of Mind (“I consider what others perceive”). These advances point toward the possibility of Artificial general intelligence (AGI), but also heighten the risk of anthropomorphizing such systems, mistaking opaque intentions for human-like cognition. To look into this, we propose the metaethics of substrate framework, which holds that the ethical challenges and responsibilities associated with intelligence cannot be reduced to abstract measures of functional performance; here, we assume that they must also account for the material substrate through which cognition emerges. In this regard, Martha Nussbaum’s capabilities approach reminds us that even biological intelligence is plural and diverse; by extension, divergence between biological-based and non-biological embodied AI may prove even greater. Recognizing these differences is vital in the current race to AGI, where market pressures often eclipse ethical reflection. Therefore, this work argues that bridging embodied AI research with frameworks such as metaethics of substrate and the capabilities approach is essential for clarifying how these systems should be situated in society, what forms of agency they may possess, and how their divergence from human cognition can be managed responsibly. es and initials are correct.