AI risks already covered by WHS laws — for now – but expect more express changes

Existing WHS and workplace laws already require employers to manage the risks posed by artificial intelligence and they cannot outsource responsibility to technology the Federal Government has confirmed, but further regulation targeting AI is likely to occur on a jurisdiction by jurisdiction basis as adoption accelerates, particularly as NSW has already amended its WHS Act to explicitly require employers to manage WHS risks from AI, algorithms and digital platforms.

Our Take

Don’t get us started on the completely unsatisfactory nature on the multiple ‘versions’ of WHS compliance we already have across multiple jurisdictions in Australia but employers just have to manage the best they can.

Our advice remains simple – if you operate in multiple jurisdictions pick the hardest compliance requirement and apply it across your business – don’t try and ‘pick winners and losers’ with multiple versions of compliance.

Whilst the Federal Government’s position is clear (i.e. AI does not sit outside work health and safety obligations) the NSW experience, and the ongoing extensive commentary on AI related matters, make it almost inevitable that other state based jurisdictions will take steps to introduce their own bespoke AI related legislation. This will be akin the steps to the approach all governments took when introducing the recent Psychosocial related changes. All were at pains to point out the existing legislation was always intended to cover both mental and physical risks but they needed to re-emphasise the mental with specific legislation because workplaces weren’t paying it enough attention.

Action Steps

• Treat AI as a WHS risk - particularly where it affects workload, monitoring, decision‑making or psychosocial health.

• Consult workers early about AI use - including how systems operate and how decisions are made or reviewed.

• Maintain human oversight - ensuring AI‑enabled decisions can be questioned or overridden where necessary.

• Review upstream obligations - particularly where software or platforms are designed, procured or customised internally.

• Prepare for change - by monitoring developments in the model WHS law review and emerging guidance on AI risks.

• Avoid complacency - as “technology‑neutral” laws still demand practical risk controls, not theoretical compliance.

Edge Legal

Relationships. Respect. Results