Scholarlink

This paper examines whether AI coaching can improve the quality, reliability, and social validity of cybersecurity response decisions in a government surveillance context, in contrast to a human-only model. Two simulations were created using an adaptive network modelling framework, one with solely human decision making and the other with AI coaching judgement without taking away agency. The models were assessed on long-term decision making, public sentiment assessment, and threat detection accuracy. Results indicate that AI coaching significantly enhances decision consistency and quality, particularly in the face of shifting public pressure. The AI-coach enabled model decreased decision unpredictability while maintaining a stronger alignment with public ideals. According to these findings, if accountability and openness are maintained, AI-guided systems have the potential to responsibly support human decision-making in government settings.

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

Using AI Coaching to Improve Cybersecurity Decision-Making: An Adaptive Network Model to Analyse AI-Supported Government Surveillance, Trust, and Threat Identification

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Artemis Hadjiloizou,
  • Jan Treur,
  • Peter H. M. P. Roelofsma

摘要

This paper examines whether AI coaching can improve the quality, reliability, and social validity of cybersecurity response decisions in a government surveillance context, in contrast to a human-only model. Two simulations were created using an adaptive network modelling framework, one with solely human decision making and the other with AI coaching judgement without taking away agency. The models were assessed on long-term decision making, public sentiment assessment, and threat detection accuracy. Results indicate that AI coaching significantly enhances decision consistency and quality, particularly in the face of shifting public pressure. The AI-coach enabled model decreased decision unpredictability while maintaining a stronger alignment with public ideals. According to these findings, if accountability and openness are maintained, AI-guided systems have the potential to responsibly support human decision-making in government settings.