Bridge has a well-established interest in exploring the role of artificial intelligence (AI) in delivering systematic literature reviews (SLRs). We have been comparing the performance of novel and powerful large language models (LLMs) against our extensive in-house datasets, which have been robustly human-QC'ed for over 100 SLRs of different types. In this first paper of the series, we share our methodology and results on AI performance in title/abstract screening.
Link to White Paper