Abstract
This protocol outlines a systematic review and meta-analysis examining the effectiveness of fully automated, AI-driven chatbot interventions in reducing subjective burden among informal caregivers. We will search 8 electronic databases (PubMed, Web of Science, Embase, CINAHL, MEDLINE, Cochrane Library, PsycINFO, Scopus) and grey literature sources from January 2010 to December 2024 for randomized controlled trials (RCTs) meeting predefined eligibility criteria. The primary outcome is caregiver burden, assessed using validated scales such as the Zarit Burden Interview. Secondary outcomes encompass caregiver mental health, quality of life, self-efficacy and care recipient status. Two reviewers will independently perform study selection, data extraction, risk of bias evaluation using Cochrane RoB 2 tool, and appraise certainty of evidence utilizing the GRADE approach. We will conduct random-effects meta-analyses, subgroup analyses, and meta-regression to compute pooled effect estimates and explore heterogeneity. If quantitative synthesis is precluded, narrative synthesis will be undertaken following SWiM guideline. Caregiver partners will provide input on interpretation and dissemination of findings.•Protocol adheres to PRISMA-P reporting standards and will be prospectively registered in PROSPERO•Graphviz code for replicating the systematic review methodology diagram is provided•Review will yield critical evidence to guide development and implementation of chatbots into caregiver support services.