The goal of the paper was to assess the safety of riding shared e-scooters and determine the causes of accidents involving them. Primary data for the research were obtained through a combination of quantitative and qualitative content analysis based on deep learning by analysing data from the WoS database entered into the VOSviewer system. A quantitative observation method was used to determine the number of shared e-scooter users who do not use safety equipment such as helmets. The data obtained were processed using the Wilcoxon test designed to test the homogeneity of two random samples in the univariate case. The observation of shared e-scooter users and social networks of e-scooter service providers showed that safety equipment is not adequately used; in addition, the users often do not know the traffic rules and often violate them. Of the 256 monitored cases of using shared e-scooters, none of the users wore a safety helmet. It can thus be concluded that it is necessary to promote the rules of the safe use of shared e-scooters. Further research could focus on proposing a methodology for assessing traffic accidents involving e-scooters and a general overview of recommendations for e-scooter safe use.