The integration of artificial intelligence (AI) into implementation of the National Environmental Policy Act (NEPA) is unfolding against a backdrop of regulatory uncertainty and heightened judicial scrutiny. Along with significant Trump Administration changes to NEPA procedures, recent court decisions have cast doubt on the Council on Environmental Quality’s authority to issue binding NEPA regulations. Meanwhile, federal agencies have initiated pilot programs to apply AI in the preparation of environmental impact statements (EISs), which raises complex questions about transparency, algorithmic bias, public participation, and whether AI-generated EISs can withstand judicial review. This Article addresses two core questions: (1) what is the appropriate regulatory pathway for governing AI in NEPA compliance—rulemaking or legislation; and (2) how can agencies ensure that AI-generated EISs meet judicial standards under NEPA and the Administrative Procedure Act? It proposes a dual-path solution: enhancing technical infrastructure through explainable, auditable AI systems, and embedding human oversight mechanisms into critical stages of the process. These reforms aim to uphold procedural integrity while responsibly harnessing AI’s potential.