Artificial intelligence is increasingly embedded in public environmental decision-making, shaping how risks are classified, resources allocated, and regulatory authority exercised. While policy attention often focuses on predictive performance and ethical principles, less scrutiny is directed toward the institutional conditions under which algorithmic outputs acquire decision relevance. This policy review addresses that gap by framing environmental artificial intelligence as decision-making infrastructure rather than as neutral analytical software. It introduces the concept of algorithmic sustainability, defined not as a technical property of algorithms but as a governance condition that aligns lifecycle environmental impacts, enforceable accountability, and procedural legitimacy. Drawing on international policy frameworks and regulatory developments, the review shows how current governance instruments insufficiently integrate lifecycle environmental footprints into decision justification. To operationalize algorithmic sustainability, the paper proposes Environmental Algorithmic Impact Assessment as a gatekeeping and renewal mechanism for artificial intelligence used in environmental governance. The review concludes that aligning algorithmic deployment with sustainability and the rule of law depends on institutional design choices made before and during system use, rather than on technical optimization alone.