Uncovering Semantic Patterns in Sustainability Research: A Systematic <scp>NLP</scp> Review

ABSTRACT This study maps how Natural Language Processing (NLP) contributes to sustainability. Using a PRISMA‐guided review of 131 English‐language articles from Web of Science (2018–2025), we combine bibliometric co‐word mapping with BERTopic to derive complementary structural and semantic views. Four themes emerge: Topic 0—Climate Change Discourse and Energy Innovation (SDG 13, SDG 7); Topic 1—Sustainability Policy and Public Engagement (SDG 11, SDG 16); Topic 2—Tech‐Driven Environmental Solutions (SDG 6, SDG 9, SDG 15); and Topic 3—Corporate Sustainability and Sustainable Infrastructure (SDG 8, SDG 9, SDG 12). Social equity remains comparatively under‐covered, with limited attention to SDG 5 and SDG 10. Theoretically, we position NLP as a bridge that supports sensemaking, coordination, and accountability across the environmental, social, and economic pillars. Practically, our synthesis highlights deployable uses of NLP for real‐time narrative monitoring, environmental, social, governance (ESG) claim assessment, and horizon scanning for emerging risks and innovation niches. For policy, coupling narrative analytics with policy‐text tracking can create feedback loops that enhance communication, legitimacy, and policy‐mix coherence for SDG implementation. By jointly leveraging network‐based and embedding‐based evidence, the review reveals thematic frontiers, makes SDG linkages explicit, and offers a topic‐based agenda with testable propositions.