Tuoya Saren (Emory University School of Law) has posted Daubert Meets NEPA: Judicial Recognition and Competence for AI-Generated EIS on SSRN. Here is the abstract:
Artificial intelligence (AI) is no longer a distant prospect in environmental law—it is already reshaping how federal agencies prepare Environmental Impact Statements (EISs) under the National Environmental Policy Act (NEPA). AI systems can process massive datasets and model environmental outcomes with speed and precision far beyond human capacity. Yet this very power exposes a new legal challenge: how can courts, built to review human reasoning, meaningfully evaluate the reasoning of algorithms that are dynamic, opaque, and often inaccessible?
This Article argues that the answer lies in adapting the principles of Daubert—the evidentiary standard used to assess the reliability of expert testimony—to the administrative review of AI-generated analyses. It proposes a reliability-based framework that translates Daubert’s core values of transparency, validation, and reproducibility into the NEPA context. Under this model, courts do not need to understand every line of code; rather, they must ensure that the agency has verified, documented, and justified how the algorithm reached its conclusions.
The Article makes three key contributions. First, it bridges the gap between evidentiary reliability and administrative accountability, showing that both can reinforce each other in AI-assisted decision-making. Second, it provides a practical roadmap for agencies to document and validate AI tools so that their use can withstand judicial scrutiny. Third, it reframes the role of the courts: deference to agency expertise is not automatic but earned through demonstrable reliability. By embedding Daubert’s logic of reliability into administrative law, the Article envisions an environmental governance system where innovation and accountability advance together—ensuring that technology serves the law, not the other way around.
Recommended.