Scholarlink

This study develops a 3D ML framework integrating microeconomics, meso-trust, and macro-markets. Using 2020–2025 global data with random forest, XGBoost, and K-means++, it analyzes impacts of AI adoption (AIAD), human-AI collaboration (HAICR), and regulation (RS) on income (RI), job loss (JL), trust (CT), and market share (AIMS). The study found that HAICR significantly inhibited the substitution effect of AI on employment (JL decreased by 0.0029), but AIAD exacerbated job loss in highly automated industries such as manufacturing (R \(^2\) = 0.576); strict supervision is the core driving factor of CT (0.060), but the effect is constrained by cultural background and tool transparency; China and the United States dominate the AI market, while India and South Korea need to strengthen infrastructure and policy adaptation; moderate supervision achieves optimal socioeconomic benefits by balancing technology iteration and employment protection. The study proposes an adaptive governance framework, recommends that enterprises prioritize the deployment of human-machine collaborative systems, implement classified supervision of policies, and build a three-dimensional model of “economy-society-user” to guide differentiated strategies. This paper provides a scalable methodology for the multi-dimensional impact assessment of AI technology, revealing the key role of scenario adaptation and dynamic policy coordination.

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

End-to-End Machine Learning for Multi-dimensional Socio-economic Impact Assessment of AI Adoption

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Na Tang,
  • Lei Mei

摘要

This study develops a 3D ML framework integrating microeconomics, meso-trust, and macro-markets. Using 2020–2025 global data with random forest, XGBoost, and K-means++, it analyzes impacts of AI adoption (AIAD), human-AI collaboration (HAICR), and regulation (RS) on income (RI), job loss (JL), trust (CT), and market share (AIMS). The study found that HAICR significantly inhibited the substitution effect of AI on employment (JL decreased by 0.0029), but AIAD exacerbated job loss in highly automated industries such as manufacturing (R \(^2\) = 0.576); strict supervision is the core driving factor of CT (0.060), but the effect is constrained by cultural background and tool transparency; China and the United States dominate the AI market, while India and South Korea need to strengthen infrastructure and policy adaptation; moderate supervision achieves optimal socioeconomic benefits by balancing technology iteration and employment protection. The study proposes an adaptive governance framework, recommends that enterprises prioritize the deployment of human-machine collaborative systems, implement classified supervision of policies, and build a three-dimensional model of “economy-society-user” to guide differentiated strategies. This paper provides a scalable methodology for the multi-dimensional impact assessment of AI technology, revealing the key role of scenario adaptation and dynamic policy coordination.