This paper presents a systematic review to identify research combining artificial intelligence (AI) algorithms with Open source intelligence (OSINT) applications and practices. Currently, there is a lack of compilation of these approaches in the research domain and similar systematic reviews do not include research that post dates the year 2019. This systematic review attempts to fill this gap by identifying recent research. The review used the preferred reporting items for systematic reviews and meta-analyses and identified 163 research articles focusing on OSINT applications leveraging AI algorithms. This systematic review outlines several research questions concerning meta-analysis of the included research and seeks to identify research limitations and future directions in this area. The review identifies that research gaps exist in the following areas: Incorporation of pre-existing OSINT tools with AI, the creation of AI-based OSINT models that apply to penetration testing, underutilisation of alternate data sources and the incorporation of dissemination functionality. The review additionally identifies future research directions in AI-based OSINT research in the following areas: Multi-lingual support, incorporation of additional data sources, improved model robustness against data poisoning, integration with live applications, real-world use, the addition of alert generation for dissemination purposes and incorporation of algorithms for use in planning.