(52z) Knowledge Base Design and Machine Learning for a Voice-Activated AI System Supporting Foreign Workers' Occupational Safety Learning and Safety Knowledge Acquisition

Foreign workers are on the rise, but there is a lack of safety education or services suitable for them. As a result, the number of related industrial accidents is also increasing, and we want to use AI speaker and voice recognized knowledge service to provide safety information related to chemicals for foreign workers with different languages.

Applications such as Google Translators can help communication, but there is a lack of support for technical terms related to the industrial safety sector. The proposed system behind the AI speaker builds industrial safety knowledge and accident cases as a knowledge base based on ontology of chemical processes, safety and hazardous materials. It creates an environment where foreign workers

can take the lead in learning and practicing safety in their own languages.

The current version of the system focuses on chemical material safety. Most knowledge comes from MSDS and the multilingual knowledge graph is implemented. Based on the user-initiated communications, the system delivers information related to chemical safety and hazards that may be involved with the hazardous substance and safety measures that should be taken in order to minimize or avoid adverse outcomes that may result from chemical exposure, hazards in the workplace, improper storage or handling of hazardous materials, chemicals in the workplace, or chemicals hazards. Depending on the operational context, the system provides employees and emergency personnel with customized chemical safety measures for handling or working with reactive chemicals, hazardous chemicals, hazardous substances, and dangerous materials in a safe and appropriate manner

Keywords: AI Speaker, MSDS, hazardous substance