Artificial intelligence has moved from the fringe to the mainstream of environment, health and safety (EHS) programs, with a new global survey finding that 82 per cent of EHS professionals report either extensive, moderate, or limited use of AI within their safety management systems.

The findings come from The Safety Shift: EHS Readiness in 2026, a study of 1,053 professionals across US industries including manufacturing, energy, and construction, conducted by Wolters Kluwer Enablon and the National Safety Council.

The report notes that organisations are moving beyond experimentation with AI and beginning to integrate it into core safety workflows — from incident reporting and root cause analysis to predictive risk modelling and the automation of compliance documentation.

However, the study also identified a significant challenge: the pace of AI adoption is outrunning governance frameworks. Safety professionals highlighted concerns around trust, accountability, and the need for human oversight when AI systems flag — or fail to flag — safety hazards.

For Australian organisations, where WHS laws place primary duties firmly on PCBUs, the question of AI governance in safety-critical decision-making is not merely theoretical. As AI-powered safety monitoring tools become standard in construction, mining and manufacturing, regulators and industry bodies will need to provide clear guidance on how these tools interact with existing duties of care.