E40 Teachers’ Perceptions of AI-Enhanced Data-Driven Decision-Making (DDDM) in Schools: A Systematic Review Existing reviews on DDDM emphasize teacher capacity, beliefs, and organizational conditions but rarely consider AI-generated data as a distinct evidentiary source in classroom decision-making (Datnow & Hubbard, 2016; Hoogland et al., 2016). Likewise, reviews of learning analytics and educational data mining focus on methodological and technical dimensions rather than on teachers' interpretive experiences (Romero & Ventura, 2020). Sociocultural approaches further highlight that data use is embedded in professional norms and institutional contexts (Biesta, 2015). Teachers' perceptions of AI-enhanced DDDM therefore reflect not only cognitive evaluations of tools but also broader negotiations of identity, ethics, and control. Purpose and Research Question Building on these perspectives, the current review seeks to clarify how teachers perceive the integration of AI into DDDM in schools. It aims to synthesize empirical evidence on both opportunities and challenges as reported across diverse contexts, while identifying underrepresented educational levels and regions. In doing so, the review bridges fragmented research strands and provides insights into responsible, teacher-centered adoption of AI in education. The review explored the following research question: What are K-12 teachers' reported perceptions (attitudes, beliefs, emotions, trust, and experiences) related to AI-enhanced DDDM in schools? Sub-questions: a) What benefits and drawbacks do teachers attribute to AI-DDDM? b) What barriers and enablers shape these perceptions and use? c) Which contextual moderators are reported? d) What ethical considerations influence teachers' trust and adoption? Methodology The review followed PRISMA-guided systematic procedures, ensuring transparent, replicable screening and synthesis of studies, minimizing bias, and establishing a reliable evidence base. Search Strategy and Eligibility Criteria The search was conducted in ERIC, Scopus, and Web of Science, with supplementary screening in Google Scholar and leading journals in educational technology, AI, and DDDM. Search terms combined four key components: K-12 teachers, DDDM, AI, and school education, using Boolean operators to capture variations in terminology across disciplines. The search process was refined through pilot testing and completed in August 2025. Duplicates and non- eligible articles were removed. Inclusion criteria focused on empirical studies in K-12 contexts examining teachers' perceptions of AI-enhanced DDDM. Theoretical, conceptual, or higher-education studies were excluded. This systematic and transparent process ensured comprehensive coverage and yielded a final dataset of 83 studies for screening and analysis. Screening and Data Extraction The screening followed a structured two-phase procedure. In the 1st phase, 83 records were reviewed by title, abstract, and keywords to assess relevance to K-12, empirical design, and AIsupported DDDM. Studies meeting predefined thematic criteria were retained, resulting in 52 eligible articles and 13 marked for re-evaluation. In the 2nd phase, full-text screening verified eligibility, yielding 25 empirical studies that fully met the inclusion criteria. For each included study, a structured data extraction table captured key information to support cross-study synthesis and comparison. The
RkJQdWJsaXNoZXIy Mjk0MjAwOQ==