This page is an educational summary for professionals working with AI. It is not the law and not legal advice. Always work from the current authoritative text linked below.
What does the code do?
It sets out how to manage psychosocial hazards at work, from identifying them through to controlling and reviewing them. The Minister approved it on 27 October 2024 under section 274 of the WHS Act 2011, and it was registered on the Federal Register of Legislation on 1 November 2024, commencing the day after registration. Its hazard list runs from job demands, fatigue and low job control through to poor organisational justice, traumatic material, remote or isolated work and intrusive surveillance.
Who does it bind?
Every duty holder under the model WHS laws as applied in the Commonwealth jurisdiction. Section 274(2) required consultation between governments, unions and employer organisations before the Minister could approve it.
What do practitioners get wrong?
Reading "code of practice" as "optional". Section 275 makes it admissible as evidence of compliance, and lets a court rely on it to decide what was reasonably practicable. Subsection 275(4) allows a different approach only where the standard is equivalent or higher.
Where does AI use touch it?
Directly, in two hazard entries. Intrusive surveillance names keyboard trackers, remote screenshots and performance-purpose GPS monitoring. Low job control names tightly scripted or machine paced work and excessive monitoring of tasks and breaks. Deploying those tools engages a named hazard. See AI-assisted psychosocial risk assessment and regulations 55A to 55D.
Bottom line
The code is evidence, not merely advice. If an AI monitoring tool matches a hazard the code names, the control expectation is already written down.
TheAICommand. Intelligence, At Your Command.
