Scholarlink

<p>Earth Observation (EO) offers valuable insights into urban environments, and integrating artificial intelligence (AI) amplifies these benefits but also brings potential risks. AI practitioners often face challenges in envisioning diverse uses and conducting thorough impact assessments of their technology, particularly for less-studied uses. To address this, we developed UrbanGen, a framework validated through studies with urban EO practitioners and compliance experts. Practitioners and experts found UrbanGen valuable both for broad thinking and reflection by listing realistic AI uses for EO (91% accuracy) and identifying under-researched uses (57% accuracy), and for in-depth thinking and decision-making by providing risk (93% accuracy) and benefit (80% accuracy) assessments. UrbanGen highlighted less-studied and upcoming uses, such as analyzing foot traffic in retail areas, monitoring environmental law compliance, and detecting crowd sizes at election rallies. While most EO uses support sustainable cities, such novel uses pose higher risks, particularly in terms of surveillance, power imbalances, and decision-making detached from on-the-ground realities. Drawing from these insights, we propose an impact assessment checklist to help the EO community maximize benefits and reduce risks from AI deployments.</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

AI in the city: impact assessment of artificial intelligence uses in earth observation

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年6月3日

查看全文
  • Sanja Šćepanović,
  • Edyta Bogucka,
  • Daniele Quercia

摘要

Earth Observation (EO) offers valuable insights into urban environments, and integrating artificial intelligence (AI) amplifies these benefits but also brings potential risks. AI practitioners often face challenges in envisioning diverse uses and conducting thorough impact assessments of their technology, particularly for less-studied uses. To address this, we developed UrbanGen, a framework validated through studies with urban EO practitioners and compliance experts. Practitioners and experts found UrbanGen valuable both for broad thinking and reflection by listing realistic AI uses for EO (91% accuracy) and identifying under-researched uses (57% accuracy), and for in-depth thinking and decision-making by providing risk (93% accuracy) and benefit (80% accuracy) assessments. UrbanGen highlighted less-studied and upcoming uses, such as analyzing foot traffic in retail areas, monitoring environmental law compliance, and detecting crowd sizes at election rallies. While most EO uses support sustainable cities, such novel uses pose higher risks, particularly in terms of surveillance, power imbalances, and decision-making detached from on-the-ground realities. Drawing from these insights, we propose an impact assessment checklist to help the EO community maximize benefits and reduce risks from AI deployments.