OBJECTIVES: AILRs refer to literature reviews undertaken with AI-based tools for one or multiple review process steps. Given the rapid and widespread adoption of AI tools in literature reviews, there is an absence of guidance from HTA bodies on their use or acceptance. This research seeks to understand if HTA processes have adapted to recognize AI.

METHODS: We reviewed relevant documents, including methodological guidance for the following HTA bodies: NICE (England), SMC (Scotland), NCPE (Ireland), HAS (France), G-BA/IQWiG (Germany), CADTH (Canada), and PBAC (Australia) to understand their acceptance of AILRs.

RESULTS: We didn’t identify any clear recommendations regarding the application of AILRs submitted as a part of the evidence package for reimbursement. NICE recommended a priority screening technique that uses a machine learning (ML) algorithm to enhance screening efficiency under its guidelines manual. This can be used to identify a higher proportion of relevant papers earlier in the screening process or to set a cut-off for manual screening. SMC refers readers to NICE methodologies. NCPE acknowledged the future of systematic reviewing via ML algorithms in their HRB-CICER report but didn’t mention anything additional. IQWiG recommended using ML-validated classifiers for identifying RCTs under bibliographic searches in general methods. Some HTAs refer to Cochrane guidance as their primary source, which itself is in the process of evaluating these algorithms to improve the efficiency of systematic review production through different initiatives, e.g., Cochrane RCT classifier, Transform project, and hybrid models, e.g., Screen4Me.

CONCLUSIONS: This research didn’t identify any explicit guidance on AILRs and indicates that much remains to be done by HTA bodies to acknowledge its potential and standardize the automation for at least the early phases of the literature review process.