Introduction The use of large language models (LLMs) in education has rapidly expanded, generating interest in their potential to support teachers through task automation and instructional planning. This review synthesizes evidence on reported changes in time devoted to educational content generation and support for planning-related task automation. Methods We conducted a systematic review of peer-reviewed literature published between 2023 and 2025, focusing primarily on secondary and higher-education contexts. Study selection followed PRISMA 2020 guidelines. Risk of bias was assessed using ROBINS-I for non-randomized studies and CASP checklists for qualitative/mixed-methods studies and for secondary evidence syntheses. Results Sixteen studies met inclusion criteria (13 primary empirical studies and 3 secondary syntheses). Across primary studies, LLM use was associated with reported time savings and perceived gains in clarity or usefulness of generated educational resources. However, outcomes and measures were heterogeneous and often self-reported, several risk-of-bias domains were rated as unclear, and evidence was concentrated in higher-education settings with small samples, limiting comparability and causal inference. Recurrent constraints included limited reproducibility, strong dependence on prompt design, and ethical or technical concerns. Discussion LLMs may support educators, but conclusions should be interpreted cautiously. Effective integration requires clear pedagogical frameworks, human oversight, and standardized evaluation in real-world settings. Systematic review registration Open Science Framework (OSF). Unique identifier: v63nj. Public URL: https://osf.io/v63nj/
Chontal et al. (Thu,) studied this question.