The construction industry has, for many years, been subject to stringent health and safety legislation for the protection of workers and the public. To ensure compliance, firms must invest a great deal in resources. However, with different legislative requirements, deadlines, and fragmentation, it is easy to overlook something or implement wrong frameworks. This study aims to investigate the applications of unsupervised machine learning (ML) on monitoring health and safety legislation and compliance on construction sites. The paper provides a systematic and comprehensive review of literature from previous studies on ML applications in construction between the years 2005-2020. A literature search from online databases was conducted using keywords. A two-step literature filtration process was used to obtain relevant publications to meet the selection criteria. The findings of the study suggest that, as technology advances shaping the future of workplace safety, ML can be used to monitor compliance and set out recommendations for future standardizations in construction. Adopting ML in the can be used to process masses of information at better speeds and accuracy to make decisions and identify anomalies that would not have been identified by humans, improving compliance. This study presents the first attempt on the applications of ML for monitoring health and safety legislation and compliance on construction sites. Future research proposes to develop a tool for contractors to use to monitor compliance.