Application of artificial intelligence in the prevention of accidents at work: a systematic review of the literature

Work accidents represent a problem not only in Brazil, but also worldwide. The International Labor Organization estimates that 2 million people die worldwide each year from work related causes. Companies, governments and workers are always looking for measures to prevent risks in the workplace. This research presents a systematic literature review, with the objective of identifying the main international publications that address the application of AI in work safety, with a focus on accident prevention. After elaborating the research protocol, and carrying out a search in the Emerald Insight, IEEE Xplore, Science Direct, Scopus and Web of Science databases, 2,369 articles were found which, after applying the exclusion criteria, 31 articles directly related to the theme were selected. . The countries with the most searches were China, the US and South Korea, with around 50% of the total. Most of the solutions presented (77%) involve work safety. Regarding the type of AI used in research, 65% use Deep Learning, while Machine Learning was used by 35%. It was evidenced that AI applied to work safety is still little explored, with a good increase from 2022.