AI Regulation Tracker  /  State law

NSW enacts new WHS duty on AI work allocation and worker surveillance; primary duties await proclamation

New South Wales has amended its work health and safety law to require employers to keep AI-driven work allocation, algorithmic management, and automated decisions safe. The core duties are enacted but do not yet bind, because they commence only on proclamation.

By Anthony Guerriero  ·  Reviewed by The Leveraged Years Editorial Desk  ·  Published July 7, 2026  ·  Last updated July 7, 2026

The New South Wales Parliament has passed the Work Health and Safety Amendment (Digital Work Systems) Act 2026, a change that pulls algorithmic management and automated workplace decisions directly into the state's work health and safety regime. The Act passed on February 12, 2026 and received assent on February 18, 2026. For employers who treated scheduling engines, productivity dashboards, and monitoring tools as neutral software, the legal frame has shifted. Those systems are now the subject of an express safety duty.

What the Act requires

The Act amends the Work Health and Safety Act 2011 (NSW). It defines a "digital work system" as an algorithm, artificial intelligence, automation or online platform, and extends the employer's primary duty of care so that worker health and safety is not put at risk by the use of such a system. A new section 21A duty addresses the specific risks that arise from how a digital work system allocates work. The definition is deliberately wide. On its face it reaches common human-resources platforms, rostering software, and performance-tracking tools, not only advanced AI models.

The Act directs employers to weigh a set of risks when a digital work system is in use. Legal analyses of the passed Act identify four categories: excessive or unreasonable workloads, excessive or unreasonable performance metrics, excessive or unreasonable monitoring or surveillance, and unlawful discriminatory practices or decision-making. Each maps to a recognised psychosocial hazard. Constant keystroke and mouse tracking, for example, sits inside the monitoring and surveillance category, and an employer that cannot justify it as reasonable, supported by a documented risk assessment, would be exposed once the duty operates.

Enacted, but not yet in force

The status matters as much as the substance. The primary duty and further WHS duty provisions do not commence on assent. They commence on proclamation, and no proclamation date had been fixed as of the middle of 2026. Some machinery provisions took effect earlier, and separate union entry provisions are set to begin only after SafeWork NSW publishes required guidelines. The Act also gives WHS entry permit holders new powers to require reasonable assistance to access and inspect a digital work system where a breach is suspected, subject to at least 48 hours' notice, which means the way these systems are configured may later be examined by a third party. Employers should confirm the live commencement position on legislation.nsw.gov.au before relying on any date, because the operative obligations are enacted but awaiting commencement.

What it does not do

The Act does not ban algorithmic management or workplace AI. It does not prohibit monitoring outright, and it does not impose a new licensing scheme on software vendors. It works through the existing WHS duty structure, which asks whether a risk has been managed so far as is reasonably practicable. The test is reasonableness and documentation, not prohibition. A keystroke logger is not unlawful in itself, but excessive or unreasonable surveillance without a recorded justification becomes a safety failure the regulator can act on.

Why NSW employers should prepare now

Several Australian law firms describe the reform as the first time a specific duty of care targeting digital work systems has been adopted in Australia. Hamilton Locke, for instance, calls it "the first time that digital work systems have been directly addressed in Australian WHS Law". That framing signals a likely model for other states, so national employers should not treat it as a purely local issue. Because psychosocial-safety duties already sit within the harmonised WHS framework used across most Australian jurisdictions, the practical groundwork, meaning risk assessments and consultation with workers, is familiar even where the digital-systems language is new.

For a US reader, the direct legal effect is confined to work performed in New South Wales. A US company with NSW operations or workers is within scope, and the Act is worth watching as an early example of workplace AI being governed through occupational-safety law rather than data-privacy or anti-discrimination statutes alone. That approach may preview how other regulators treat algorithmic management.

The immediate task is documentation. Employers should build an inventory of systems that allocate work, set metrics, or monitor staff, assess the psychosocial risk each creates, consult affected workers, and record the reasoning. Doing that before proclamation converts a pending obligation into a defensible position.

Browse the full AI Regulation News tracker

Informational analysis for working professionals, not legal advice. Confirm how any rule applies to your situation with qualified counsel.