The contribution of ontologies to the creation of knowledge bases for the management of legal compliance in occupational health and safety

International audienceOccupational health and safety (OHS) is at the heart of corporate policies and the increasing trend towards litigation has contributed to the inflated number of regulatory texts published each year. OHS professionals are therefore faced with the need to process a growing amount of data in order to remain in compliance. They train themselves in law, draw upon internal expertise (the company's legal department) or external consultants (advisory and enforcement agencies) and equip themselves with information systems. These systems are mainly 'judicial monitoring' databases that contain regulatory texts and associated comments. For the most part, this data is complex to interpret and difficult to exploit given the growing number of texts, the expertise needed to understand them and the difficulty in being sure that they are relevant to the activities of the company. This need for clarity may be met by models and methods from artificial intelligence. Amongst these, we have selected the concepts of the ontology and the knowledge base. An ontology is a model of a particular domain of knowledge created for a specific purpose. Mainly used for the formalization of knowledge, it can also be used to define a common language between an information system and a community of actors (i.e. OHS professionals). A knowledge base is a set of knowledge that is specific to a given domain that can be exploited by a technology. The coupling between the ontology and the knowledge base makes it possible to automate certain tasks, such as monitoring, auditing, ensuring compliance, etc. This three-part article describes a methodology and an associated software tool that can ensure compliance. The first part identifies elements used in the definition and the proven benefits of ontologies and knowledge bases. The second describes the approach taken in the design of the ontology and its implementation as a rule base using a bespoke knowledge base editor. Finally, a concrete example of the formalization of regulatory knowledge (using the theme of harassment and violence at work) is presented to demonstrate the contributions and the limitations of the tool