Üretken yapay zekâ destekli karar verme sistemleri: PRISMA 2020 akışına dayalı sistematik literatür incelemesi ve tematik sentez

Counts differ because each database indexes a different set of publications. We treat OpenAlex as the canonical count; Google Scholar is not shown (no API, and crawling it violates its ToS).

This study systematically reviews the literature on generative artificial intelligence (GenAI)- and large language model-supported decision-making systems, focusing on research trends, application domains, methodological characteristics, risks, and governance approaches. A Web of Science and Scopus search conducted on May 13, 2026, covering 2020–2026, was reported in accordance with PRISMA 2020. After duplicate removal and eligibility assessment, 31 of 402 records were included in the synthesis. Studies were appraised using MMAT 2018 or design-appropriate JBI checklists; across 249 items, 171 “Yes,” 33 “No,” and 45 “Unclear” judgments were recorded, without calculating an overall score or quality category. The thematic synthesis identified six themes: decision-support capacity; trust and user acceptance; explainability and uncertainty management; human oversight and governance; risks and responsible use; and decision architecture and multi-criteria decision-making. Based on these findings, a six-component Integrative Conceptual Framework for Generative AI-Supported Decision-Making was proposed. The findings indicate that GenAI should function as an explainable, auditable, human-supervised, and institutionally governed decision-support component, while human decision-makers retain authority and responsibility.