Objective The objective of this presentation is to provide an introduction to the symposium entitled: ‘A road map for artificial intelligence and occupational health in the 21st century.’

Methods There are many important applications for industry and occupation codes in occupational hygiene and epidemiology. Many health studies collect job-related information in the form of free text (regarding a job title or tasks normally performed by a worker) which needs to be classified into a code in order to group similar workers together for analysis. Additionally, one of the standard tools for assigning exposure metrics to individuals in large studies is a job-exposure matrix (JEM) which is also linked to an industry code, an occupation code, and often both.

Results For the question of coding free-text into an appropriate code, for most of the history of occupational epidemiology, this has been done manually (and ideally by an expert coder(s)). This can be extremely time-consuming and has its own issues of reliability. For the question of JEMs, if they are created in one coding system, it can be difficult and subjective to re-code them to a system for use in a different country or year, for example (many coding systems are refined over time). There are many new and promising AI and/or machine learning applications and tools being developed around the world.

Conclusions This introductory talk will set the stage for the symposium, providing a background on the issues with occupational coding systems and code assignment in general, and how automatic coding systems could improve occupational epidemiology.