The regulator's new exposure count: two in three workers probably exposed, and one in 20 at or around the silica exposure standard

What workers are exposed to, by the regulator’s count

The survey asked 6,617 workers aged 15 and over what they did in their last working week, then ran those answers through a set of expert rules that estimate whether, and how much, each task exposes a worker to each of 40 agents: dusts, metals, chemicals, solvents, radiation, products of combustion, heat and night shift work. Results are weighted to the Australian workforce. Every percentage below is a share of all workers unless it says otherwise.

Three things in that chart need reading carefully. First, solvents at 31.2 per cent is almost all benzene, at 31.0 per cent of all workers, and most of that, 25.4 per cent, is assessed as low level. Second, radiation is mostly solar UV (32.4 per cent) and ocular UV (32.6 per cent), which overlap heavily: 91.6 per cent of workers exposed to solar UV are also exposed to ocular UV. Third, the dusts bar is short but half of it is dark. Silica (13.6 per cent) and wood dust (8.2 per cent) are the main dust exposures, and wood dust is more often high level (5.2 per cent) than any other level.

“High” has a precise meaning here. For airborne agents, the report’s methodology defines a high-level probable exposure as exposure at or around the WES, the workplace exposure standard. Medium is roughly 10 to 80 per cent of the standard, and low is above the general community but under about 10 per cent. For heat and radiation the rules use the equivalent reference, such as the heat guidance or the annual radiation limits.

The high-risk agents, and who carries them

The report singles out agents linked to cancer or chronic disease. For each one it gives the share of all workers, a national estimate, and where exposure concentrates. The industry column is the share of that industry’s workers; the last column is the share of apprentices and trainees.

Every figure is as the report states it. Counts are the report’s own extrapolation to about 15.0 million employed people (ABS Labour Account, December quarter 2025); we have not made any.

Source: 2026 AWES Analytical Report, Table 3 (pp. 6 to 8) and the high-risk agent sections, pp. 34 to 49. The report notes the mining figure for asbestos may reflect naturally occurring asbestos.

Silica, in detail

Silica is the agent where breadth and intensity meet. Of all workers, 6.2 per cent are assessed at low level, 2.4 per cent at medium and 5.0 per cent at high, so more than a third of silica exposure is assessed at or around the standard. By industry it is 71.8 per cent of construction workers and 63.8 per cent of mining workers, then agriculture, forestry and fishing (30.8 per cent) and electricity, gas, water and waste services (22.4 per cent). By occupation it is 89.9 per cent of construction and mining labourers and 86.2 per cent of construction trades workers. Men are far more likely to be exposed than women, 22.6 per cent against 4.2 per cent, and workers outside capital cities more than those in them, 17.6 against 11.3.

The apprentice figure is the one that stands out: 35.7 per cent of apprentices and trainees are estimated to be exposed to silica, against 12.5 per cent of other workers. The same gap shows for heat (51.1 against 28.8 per cent), lead (17.0 against 3.1) and asbestos (13.7 against 2.0).

Many hazards at once

Exposure clusters. Across all workers, 49.4 per cent are probably exposed to more than one agent and 22.2 per cent to five or more. By industry, five or more is the estimate for 68.2 per cent of construction workers, 60.1 per cent in agriculture, forestry and fishing and 60.0 per cent in mining, against 1.8 per cent in financial and insurance services. By occupation it is 82.1 per cent of construction trades workers and 81.4 per cent of construction and mining labourers.

One finding is not a chemical at all. Environmental tobacco smoke, including vapes, reaches 25.8 per cent of all workers and is in the top ten exposures for every industry, from which the report concludes there may be gaps in the enforcement of smoking restrictions across work environments in Australia.

The survey also measured psychosocial hazards by a different method, asking workers how often things happen. 68.3 per cent say their job requires a lot of concentration often or all of the time and 37.3 per cent work under tight deadlines or pressure that often; 21.8 per cent disagree that they have a say in changes at work that affect them. The 68.3 is a different measure from the 68.0 above and the two should not be read together.

What it does not tell you

The report is plain about its limits. The analytical report says This study has not obtained measurements of actual exposure. Every figure is a probable exposure inferred from tasks. It covers the last working week in the main job, not lifetime exposure.

It is also not a verdict on any one workplace. In the report’s words, The findings are not a detailed assessment of risk in individual workplaces.

On controls, the survey does not report which controls workplaces use. What it does say is that its rules can take account of protective measures such as closed systems and ventilation, and of respiratory or eye protection, depending on the task. So for some tasks a high-level estimate is the exposure that remains after the controls the worker described, not the hazard before them.

Fieldwork ran from 2 December 2025 to 18 February 2026, in summer. The report flags the timing as possibly inflating solar UV exposure, and makes the same caveat for heat.

And it cannot tell you whether anything is getting better or worse. The report makes no comparison with earlier Australian exposure work: it names the earlier Australian Work Exposures Study only as a previous user of the same assessment system, and calls the 2026 survey a baseline for monitoring change. Its only comparisons are international and it cautions that those use different methods: New Zealand’s carcinogens survey put silica exposure at 10.3 per cent of workers, against 13.6 here.

Our read

This section is our opinion, built on the figures above.

Read the high-level column, not the headline. The 68.0 per cent is dominated by sun, low-level benzene, diesel and second-hand smoke. The number a duty holder should hold against their own work is the 28.9 per cent probably exposed at a high level, and within it the agents where the dark bar is a large share of the light one: silica, wood dust and formaldehyde (2.3 of its 6.5 per cent is high level).

For silica, benzene and formaldehyde, “high” is pegged to a number the regulator does not stand behind. The survey defines high as at or around the standard. For those three, the standard is one of nine that WHS ministers did not agree to change: Safe Work Australia said on 24 June 2026 there will be no changes to the limits for the nine from 1 December 2026, and its own Decision Regulation Impact Statement says evaluation of the current standard for each of the nine has determined they are not health-based. We covered that decision in our piece on the nine chemicals. Put the two documents side by side and a medium-level silica estimate, 10 to 80 per cent of a standard its own evaluator calls not health-based, is not a reason to stop at the risk assessment. In our view it is the case for air monitoring under regulation 50, which applies where a PCBU is not certain on reasonable grounds whether the standard is exceeded.

The newest worker is the most exposed one. The survey does not say why apprentices and trainees show nearly three times the silica exposure of other workers. Our inference is that the dustiest tasks may be falling to the least experienced people. A risk assessment that looks at the task but not at who is doing it would miss that, and the survey’s numbers are the reason to ask.

This survey counts exposure; it cannot count monitoring. Safe Work Australia’s own review of health and air monitoring conceded in August that the actual rate of monitoring is unclear. Now there is a national estimate of who is exposed, and still no national picture of who is being watched. Matching the two is the obvious next step, and it is work only the regulators can do.