Scholarlink

<p>This paper examines the paperclip apocalypse concern for artificial general intelligence. This arises when a superintelligent AGI with a simple goal (ie, producing paperclips) accumulates power so that all resources are devoted towards that goal and are unavailable for any other use. Conditions are provided under which a paperclip apocalypse can arise, but the model also shows that under certain architectures for recursive self-improvement of AGIs, a paperclip-maximising AGI may refrain from allowing power capabilities to be developed. The reason is that such developments pose the same control problem for the AGI as they do for humans (over AGIs) and hence, threaten to deprive it of resources for its primary goal. <i>Journal of Economic Literature</i> Classification Numbers: D50, D80.</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

Self-regulating artificial general intelligence

  • 期刊论文

  • 发表日期：

    2025年12月1日

查看全文
  • Joshua Gans

摘要

This paper examines the paperclip apocalypse concern for artificial general intelligence. This arises when a superintelligent AGI with a simple goal (ie, producing paperclips) accumulates power so that all resources are devoted towards that goal and are unavailable for any other use. Conditions are provided under which a paperclip apocalypse can arise, but the model also shows that under certain architectures for recursive self-improvement of AGIs, a paperclip-maximising AGI may refrain from allowing power capabilities to be developed. The reason is that such developments pose the same control problem for the AGI as they do for humans (over AGIs) and hence, threaten to deprive it of resources for its primary goal. Journal of Economic Literature Classification Numbers: D50, D80.