A consolidated framework for implementation research informed perspective on AI implementation in special education: a systematic synthesis of evidence

Background Integrating AI into special education for students with intellectual disabilities faces ongoing non-technological challenges like a lack of support, poor context fit, and infrastructure issues, with limited research addressing these. A comprehensive framework for these challenges is missing.Methods This systematic review uses a consolidated framework for implementation research (CFIR) to analyze eight AI intervention trials in real-world special education settings. Studies include randomized controlled trials (RCT), and single-case designs. Three additional teacher perception surveys are analyzed separately as supplementary evidence (Supplementary File 2). Implementation factors were categorized into CFIR domains, thematically analyzed, and assessed with GRADE-CERQual.Results The review identified four overarching themes, three assessed at moderate and one at low-to-moderate confidence (GRADE-CERQual): (1) human mediators (teachers, therapists, facilitators) are active, not peripheral; (2) a trade-off exists between engagement and accessibility; (3) multi-level contextual fit (user, setting, systemic) is crucial; (4) tech development moves from static tools to data-informed, adaptive platforms, still emerging. Facilitators include collaborative design and training; barriers involve misalignment with classroom realities, resource gaps, and weak infrastructure.Conclusions Successful integration requires a shift from a tech-focused approach to an implementation-focused one, emphasizing support systems and systemic readiness. CFIR helps stakeholders bridge research and practice for equitable, sustainable outcomes.