Facilitating public scrutiny of EIA reports with open data and artificial intelligence: insights from a Mexican case study

Public scrutiny is a cornerstone of Environmental Impact Assessment (EIA), yet their technical complexity methods often limits the ability of stakeholders to meaningfully review EIA reports. This paper introduces a practical approach that integrates Open Data (OD), Artificial Intelligence (AI), and comparative analysis of impact significance assessment to evaluate whether key sources of uncertainty are adequately addressed in EIA reports. We apply this approach to the case of the, focusing on a technical-scientific review of an EIA report of Tren Maya, commissioned by Mexican environmental authorities. We demonstrate how our approach can be employed to assess epistemic and ontological uncertainty in biodiversity impact assessments. This approach can identify weaknesses in expert judgments, data quality, and adherence to core EIA principles. Results indicate that the EIA report underestimated biodiversity impacts and exhibited methodological inconsistencies in indicator scoring and aggregation. In general, OD-AI tools supports evidence-based scrutiny under procedural constraints, while comparative analysis highlights the risks of biased or inconsistent judgments in conventional EIA practice. The approach is transparent, operationally accessible, and adaptable to other EIA domains and contexts. OD-AI supports public scrutiny of EIA reports.OD enables efficient appraisal of biodiversity baseline data.AI enables traceable, literature-backed assessments of impact significance.Comparative analysis evaluates compliance of expert judgment with core EIA principles. OD-AI supports public scrutiny of EIA reports. OD enables efficient appraisal of biodiversity baseline data. AI enables traceable, literature-backed assessments of impact significance. Comparative analysis evaluates compliance of expert judgment with core EIA principles.

公众监督是环境影响评价（Environmental Impact Assessment, EIA）的核心基石，但其技术复杂度较高的方法往往会限制利益相关方对环境影响评价报告开展有效审查的能力。本文提出了一种融合开放数据（Open Data, OD）、人工智能（Artificial Intelligence, AI）以及影响显著性评价比较分析的实用方法，用于评估环境影响评价报告是否充分处理了关键不确定性来源。我们将该方法应用于相关案例，聚焦于由墨西哥环境主管部门委托编制的特伦玛雅（Tren Maya）铁路环境影响评价报告的技术科学审查工作。我们展示了该方法可用于评估生物多样性影响评价中的认知不确定性与本体论不确定性。该方法能够识别专家判断、数据质量以及核心环境影响评价原则遵守情况方面的不足。研究结果显示，该环境影响评价报告低估了生物多样性影响，且在指标评分与指标聚合环节存在方法学不一致性问题。总体而言，开放数据-人工智能（OD-AI）工具能够在程序约束下支持循证审查，而比较分析则凸显了传统环境影响评价实践中判断存在偏差或不一致的风险。该方法具备透明度高、操作便捷且可适配其他环境影响评价领域与应用场景的优势。开放数据-人工智能（OD-AI）支持对环境影响评价报告的公众监督。开放数据可高效评估生物多样性基线数据。人工智能可开展可溯源、有文献支撑的影响显著性评价。比较分析可评估专家判断是否符合核心环境影响评价原则。开放数据-人工智能（OD-AI）支持对环境影响评价报告的公众监督。开放数据可高效评估生物多样性基线数据。人工智能可开展可溯源、有文献支撑的影响显著性评价。比较分析可评估专家判断是否符合核心环境影响评价原则。