Safety is critical in any work environment, and this initiative demonstrates the ability of a deep-learning-based framework to provide real-time feedback to on-site supervisors, playing a pivotal role in promptly identifying and rectifying deficiencies. Artificial intelligence (AI) has extensive implications, particularly within industrial environments where safety is paramount. This study delves into the complex deep-learning network through real-time video analysis, utilizing AI to evaluate the use of personal protective equipment (PPE) in workplaces, with a key emphasis on adherence to Occupational Safety and Health Administration (OSHA) guidelines. By leveraging the YOLOv6 algorithm, the AI-based system monitors activities, learns, and adjusts, providing a dynamic framework capable of addressing the intricacies of safety regulations. Tasks such as preprocessing, normalization, and noise reduction are crucial for maintaining the integrity and relevance of the data. The system demonstrated an impressive accuracy rate of 88.5% when processing real-time data from the camera, ensuring its reliability in practical applications. The YOLOv6 model has been trained to utilize nine evenly distributed anchor boxes across three scales, enabling precise object detection and classification with high accuracy. A significant challenge encountered during the project was the latency observed in processing real-time feeds, underscoring the need for optimization. Furthermore, while the system exhibited remarkable precision in controlled environments, real-world obstacles such as dense frames, fluctuating lighting conditions, and obstructions highlighted areas for future enhancements.