Data-Driven Safety Messaging in Mining and Processing Facilities: Effects on Incident Rates and Reporting Behavior

This paper focuses on the impact of data-based safety messages and their relationship to processing and mining operations compared to the level of incidents and incident-reporting behavior. The archival safety histories of the many sites denote that three major elements of safety messaging are to be taken into consideration, that is, channel composition, message frequency, and readability. These results show that fewer incidents occurred through near-miss reporting regularly and with consistent communication schedules that improved the number of incidents multi-dimensionally through digital-offline methodologies. Readability was observed to play the key role; optimized messages, which were simplified and optimized using AI, improved understanding and interaction and risk patterns concealed in natural language processing. These results generalize the healthcare and manufacturing industry models of transplant safety climate and incident reporting in the mining sector and empirically demonstrate that AI-driven surveillance can potentially broadcast the messages of risk. The theoretical observations presuppose the propensity to introduce facility-wide platforms of communication and create a facility-specific playbook that would facilitate the personalization of delivering messages. Future research ought to be conducted on longitudinal AI-enabled experiments within high-risk industries in a bid to enhance cross-industry generalizability.