Scholarlink

<p>As artificial intelligence (AI) is increasingly integrated into welfare governance, concerns about its legitimacy and public support are emerging. Across four studies with Chinese participants (<i>N</i> = 1336), this research examines how decision-maker identity (AI vs. human) influences public evaluations of legitimacy and program support, and explores the mechanisms underlying these effects. Studies 1 and 2 indicated that when outcome information was absent, legitimacy judgments of AI and humans were similar, although support slightly favored human-administered programs. These differences were associated with perceptions of the decision-maker’s capacity to bear forward-looking moral and social responsibilities, rather than with perceived fairness or experience. However, these identity-based differences were not robust. They did not remain significant under conservative Holm–Bonferroni corrections and were attenuated when direct outcome (Study 3) or side-effect information (Study 4) was provided. Individual beliefs in machine heuristics and justice also moderated responses to AI and human decision-makers. These findings suggest that public responses to AI in welfare governance are conditional, rather than reflecting stable preferences or generalized resistance. They further highlight the role of perceived responsibility-bearing capacity in shaping acceptance and support in decision-making scenarios.</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

Outcomes Override Identity: How Forward-Looking Responsibility and Outcome Valence Shape Legitimacy and Support in AI Welfare Systems

  • 期刊论文

  • 发表日期：

    2026年3月8日

查看全文
  • Caifeng Xie,
  • Xiaojun Ding,
  • Minqiang Xu,
  • Chengcheng Jiao,
  • Sirui Fu,
  • Jiayi Xin,
  • Feng Yu

摘要

As artificial intelligence (AI) is increasingly integrated into welfare governance, concerns about its legitimacy and public support are emerging. Across four studies with Chinese participants (N = 1336), this research examines how decision-maker identity (AI vs. human) influences public evaluations of legitimacy and program support, and explores the mechanisms underlying these effects. Studies 1 and 2 indicated that when outcome information was absent, legitimacy judgments of AI and humans were similar, although support slightly favored human-administered programs. These differences were associated with perceptions of the decision-maker’s capacity to bear forward-looking moral and social responsibilities, rather than with perceived fairness or experience. However, these identity-based differences were not robust. They did not remain significant under conservative Holm–Bonferroni corrections and were attenuated when direct outcome (Study 3) or side-effect information (Study 4) was provided. Individual beliefs in machine heuristics and justice also moderated responses to AI and human decision-makers. These findings suggest that public responses to AI in welfare governance are conditional, rather than reflecting stable preferences or generalized resistance. They further highlight the role of perceived responsibility-bearing capacity in shaping acceptance and support in decision-making scenarios.