Scholarlink

<p>A growing literature discusses the potential benefits and risks of advanced artificial intelligence, yet relatively little work has empirically examined how laypeople evaluate non-human intelligences when they are presented as potential future threats. Using an online survey of 721 UK adults, we examined responses to vignettes describing either an artificial intelligence system or an alien intelligence that could pose a threat to humanity. Participants reported greater fear and judged pre-emptive destruction to be more morally permissible in the artificial intelligence condition than in the alien intelligence condition. By contrast, differences in intelligence level (superhuman versus subhuman) played a limited role in shaping responses. Perceived friendliness during interactions systematically reduced fear, whereas perceived unfriendliness increased it across conditions. These findings suggest that judgments of non-human intelligences in threat contexts are driven more by socially salient category labels and framing than by specified levels of intelligence or capability. The results have implications for understanding how public moral judgments about emerging technologies may be shaped by narratives and labels rather than by their technical properties.</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

People fear and find it morally permissible to pre-emptively destroy A.I. more than aliens in hypothetical threat scenarios

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年5月20日

查看全文
  • Jamy Li

摘要

A growing literature discusses the potential benefits and risks of advanced artificial intelligence, yet relatively little work has empirically examined how laypeople evaluate non-human intelligences when they are presented as potential future threats. Using an online survey of 721 UK adults, we examined responses to vignettes describing either an artificial intelligence system or an alien intelligence that could pose a threat to humanity. Participants reported greater fear and judged pre-emptive destruction to be more morally permissible in the artificial intelligence condition than in the alien intelligence condition. By contrast, differences in intelligence level (superhuman versus subhuman) played a limited role in shaping responses. Perceived friendliness during interactions systematically reduced fear, whereas perceived unfriendliness increased it across conditions. These findings suggest that judgments of non-human intelligences in threat contexts are driven more by socially salient category labels and framing than by specified levels of intelligence or capability. The results have implications for understanding how public moral judgments about emerging technologies may be shaped by narratives and labels rather than by their technical properties.