Along with the growing incidence of cancers worldwide, the volume of related research is expanding at an unprecedented pace, leading to thousands of new cancer studies published each year [1]. These developments highlight the need for comprehensive reviews, including rapid reviews and network meta-analysis, with a particular emphasis on systematic reviews (SRs). SRs are essential for evidence-based medicine and serve as resources to inform and guide clinicians and researchers [2]. However, the completion of SRs is time-consuming. On average, a SR takes approximately 67.3 weeks to complete at well-established research institutions from protocol initiation to submission for publication [3], [4]. Among the steps to conduct a new SR, screening literature search results is one of the most time-consuming steps, which includes two stages: title and abstract screening stage and full-text screening stage [5], [6].

Given the increasing demand for SRs and the constraints associated with their production, there has been a growing interest in leveraging artificial intelligence (AI)-based tools to expedite the process. In recent years, a variety of AI-based automation tools have emerged with capabilities to automate, either fully or partially, critical steps in the SR process, including screening, literature search, data extraction, and synthesis, etc. [7], [8], [9]. While several review articles have explored these automation technologies, they did not focus specifically on their application for SRs in cancer research. For instance, a scoping review published in 2022 evaluated nine AI-based automation tools but did not delve into their specific applicability or effectiveness for SRs of primary studies on cancer [10]. Additionally, these publications indicate that the performance of one AI tool can be varied across different datasets of different diseases. Thus, this SR aims to address this gap to guide future decision-making regarding the use of AI in SRs, potentially leading to improved efficiency in the production of SRs in cancer research. Thus, the research question of this SR is: In SRs focusing on cancer research, which AI-based automation tool demonstrates the best accuracy performance and workload savings for the literature screening step, compared with the reference standard of human screening?