Automatic completion of COSHH risk assessment forms using semantic representation of GHS and CLP regulation

Abstract

The completion of a risk assessment form is a sequence of three events: chemical information discovery and acquisition; chemical information manipulation; and, assessment form template interpolation. For organisations, the costs associated with manual risk assessment are very significant; ranging from the wages of those who complete and review the forms, to the damages paid to those who are affected by any errors. We present a software system that uses Semantic Web technologies to assist with and automate the completion of COSHH risk assessment forms. We use an RDF representation of the GHS and CLP regulation to describe the hazards and precautionary statements that are associated with a library of chemical substances. Aggregating the RDF data for a collection of chemical substances enables the automatic generation of a ready-to-use, human- and machine-readable risk assessment form. The software system is backed by an RDF triple-store and SPARQL endpoint, and fronted by a lightweight, mobile-friendly Web application. We find that the introduction of software systems for automated risk assessment is, on the whole, beneficial to organisations. However, we also encounter a number of new risks, including: the risk that users obtain and utilise the “correct” answer to an “incorrect” question; the risk that users implicitly trust and/or rely upon the software system; and, the risk that the data gathered by the software system could be used to infer the research agendas of both individual users and organisations as a whole. Finally, we report on recent work to implement our software system using the Sigma-Aldrich catalogue as the primary data source

More information

Identifiers

ORCID for Jeremy G. Frey: orcid.org/0000-0003-0842-4302

Catalogue record

Export record