Glad You Asked 6 - Do WHS laws apply to the AI and software we use?
The state that regulated the algorithm
In February 2026, NSW became the first state to pass laws making it explicit: if your business uses digital work systems, including AI, algorithms and automation, you must make sure they don't put workers' health and safety at risk. The laws call out unsafe workloads, unreasonable performance tracking, excessive surveillance and discriminatory work allocation by name.
Sit with that for a second. Your rostering app, your productivity dashboard, your AI tools. All of it is now a safety conversation. And here's the part that matters wherever you operate: NSW didn't invent this duty, it wrote it down. Your duty of care has always covered your systems of work, and those systems are increasingly digital. What's new is that regulators are saying it out loud.
“But the software did it”
This line won't save you.
When an automated roster gives someone a late finish and an early start three days running and they make a mistake on fatigue, that's your hazard. When software quietly ratchets picking targets up and people start rushing lifts to keep their numbers green, that's your hazard. When a monitoring dashboard means nobody takes a proper break because idle time looks bad, that's your hazard too.
You chose the system. You configured it, or didn't. The duty sits with you, not the vendor and not the algorithm.
It’s not just the mental load
Picture a worker moving through a warehouse, eyes on a scanner, pick instructions coming through a headset, attention locked on the next task the system feeds them. Now put them in live forklift traffic. The technology is performing exactly as designed. But the traffic management plan was written before anyone was walking the floor with their head in a device. Digital systems change how people move, look and pay attention, and the physical risks around them need to be reassessed to match.
The rollout nobody was asked about
The pattern we keep seeing: a business switches on a new tool that allocates work, sets targets or monitors performance, and the workforce finds out when it lands. Two problems. The law requires you to consult workers before changes that affect their health and safety, and a system that decides who does what absolutely qualifies. And if nobody in the business can explain how the tool makes its decisions, that's a red flag, not an excuse.
What good looks like
You don't need to rip anything out. You need to treat your digital systems like any other part of the workplace.
Know what you're running: list every system that allocates work, sets targets or monitors people.
Set the guardrails: break rules, shift caps, reasonable targets, not the default settings built to maximise output.
Consult before you switch anything on.
Keep a human in the loop on decisions the system makes about people.
There is a way to get ahead of this. We're working with businesses to map the digital systems they actually rely on, test what those systems are doing to workloads, monitoring and work allocation, consult the workers who use them, and land on controls that fit. Built around how their business runs, not a template.
So do WHS laws apply to the AI and software you use? Yes. If a system decides how people work, how hard they work, or how closely they're watched, the risk it creates is yours to manage.