Scholarlink

<p>This research investigates how AI characteristics (autonomy, interactivity, ambidexterity, and techno-bias) influence decision-making in AI-powered hiring systems, with resource allocation as a moderator. Grounded in the next-generation PCI framework and resource-based view, the study analyzes e-survey data from 176 professionals. Findings show that autonomy, interactivity, and ambidexterity positively relate to decision-making, while techno-bias exhibits an inverted U-shaped relationship: Moderate levels enhance decisions, whereas very low or excessive levels impair them. Resource allocation significantly moderates several of these relationships. This research introduces techno-bias as an inherent AI characteristic arising from embedded human biases, demonstrates that moderate bias can be functionally tolerable, and provides a preliminary graph to calibrate ideal levels of bias for optimal decision-making rather than merely mitigate or eradicate it.</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

Techno-bias as an AI characteristic: calibrating bias for optimal decision-making

  • 期刊论文

  • 发表日期：

    2026年6月17日

查看全文
  • Helmi Issa,
  • Elisabeth Gressieux

摘要

This research investigates how AI characteristics (autonomy, interactivity, ambidexterity, and techno-bias) influence decision-making in AI-powered hiring systems, with resource allocation as a moderator. Grounded in the next-generation PCI framework and resource-based view, the study analyzes e-survey data from 176 professionals. Findings show that autonomy, interactivity, and ambidexterity positively relate to decision-making, while techno-bias exhibits an inverted U-shaped relationship: Moderate levels enhance decisions, whereas very low or excessive levels impair them. Resource allocation significantly moderates several of these relationships. This research introduces techno-bias as an inherent AI characteristic arising from embedded human biases, demonstrates that moderate bias can be functionally tolerable, and provides a preliminary graph to calibrate ideal levels of bias for optimal decision-making rather than merely mitigate or eradicate it.