AI spend grows but Australian workers doubt safety gains

Australian employers are spending more on digital safety systems and artificial intelligence, but many workers say the changes are not making day-to-day conditions safer, according to new research commissioned by safety and compliance platform provider Rapid Global.

The Workplace Safety Australian Market Research Report draws on a survey of more than 1,000 safety managers, workers, and contractors in high-risk industries. It suggests a widening gap between formal safety processes and workers' confidence on site.

While 65% of workers said safety processes were clear and practical, only 41% believed safety was taken seriously by everyone, all the time. The report described that gap as a risk factor that can weaken trust and leave hazards unaddressed.

Culture gap

Professor Dr Andrew Sharman, CEO of the International Institute of Leadership & Safety Culture, linked the results to broader patterns across industries.

"Safety is often well documented, yet not consistently felt by people on the ground," Sharman said.

"Bridging the gap between policy and practice is less about systems alone and much more about leadership. Trust is the critical differentiator."

Even as organisations add more tools, the research found many still struggle with usability and execution. Two-thirds of managers (67%) said unifying safety systems would simplify compliance, placing a high value on platform integration.

However, fewer than half of executives (45%) said their current safety tools used modern, easy-to-use technology. Among frontline workers, 41% described safety software as easy to understand. That fell to 30% in larger organisations, suggesting scale and complexity can make digital systems harder to use.

The report warned that when tools are difficult to use, workers may resort to manual workarounds or avoid reporting. That can create blind spots for employers and reduce the value of data produced by digital systems.

Paper persists

Despite years of digitisation, 23% of Australian businesses still rely on paper for critical safety tasks. The report linked this to industries with remote worksites and large contractor workforces, where connectivity and system access can be inconsistent.

Friction also persists in organisations that have deployed digital tools. Half of managers said incident reporting could be easier. Nearly one in four workers said they had personally seen incidents go unreported, indicating barriers at the point of capture and escalation.

The research also highlighted weaknesses in enforcement at site entry. More than one-third of workers and managers said it was still possible to enter a workplace even when training was incomplete or expired. A majority of managers (62%) said linking induction data directly to site access systems would improve safety.

AI fault line

Artificial intelligence emerged as a dividing line between management expectations and worker sentiment. Only 12% of managers said they saw no value in AI for workplace safety, but views on its impact varied sharply by role.

Nearly two-thirds of managers (64%) believed AI and robotics would fundamentally transform workplace safety within five years. Only 25% of workers agreed, suggesting scepticism about how AI will affect daily work and whether it will address immediate hazards.

Governance also appears to be lagging interest. The research found 41% of managers were already using AI tools not officially provided by their organisation for safety tasks. That points to experimentation outside policy and procurement, with potential implications for data quality, accountability, and consistency.

Managers also expressed caution about handing decision-making to automated systems. Six in ten said AI should support data analysis rather than make safety decisions, reinforcing the role of human judgement in higher-risk environments and complex operational settings.

Adoption differed by state. Organisations in New South Wales, Queensland, and Western Australia were more than twice as likely to use AI-enabled safety systems as those in Victoria and South Australia, according to the survey.

Complexity risk

Ezequiel Gonzalez, Head of Revenue at Rapid Global, said the findings point to safety risk shaped by complexity rather than intent.

"Workplace safety in Australia has come a long way, but there is no room for complacency," Gonzalez said.

"Complex, high-risk environments require more than ticking boxes. Technology should not replace human judgement but make it sharper. When systems are easier to use and data is easier to act on, safer outcomes follow."

The report argued that improvement correlates less with the number of tools deployed and more with reducing friction for users and automating enforcement. It said organisations are more likely to see better outcomes when safe behaviour is the easiest option in daily routines, particularly for contractors and dispersed workforces.

With many employers reviewing platform integration, data quality, and rules around AI use, the findings suggest the next phase of safety investment may focus on usability, compliance enforcement at entry, and tighter governance of emerging technologies.