Australia: Is AI raising the bar on your WHS obligations?

Artificial intelligence (AI) is being used in Australian workplaces now, allocating tasks, monitoring fatigue, detecting hazards and influencing decisions that directly affect worker safety. The legal frameworks governing workplace safety (developed long before AI) were not designed with AI in mind and are now playing catch-up. For officers and corporate duty holders, this raises two distinct but equally pressing questions:

• How do you manage the WHS risks introduced by AI?     
• Could you be required to implement AI in order to discharge your existing WHS obligations?

AI as a hazard

AI adoption introduces new risks that must be managed under existing WHS laws. The use of AI to track workers, manage performance and allocate tasks can create significant psychosocial hazards, including low job control, poor organisational justice, and excessive monitoring, all of which are well-recognised WHS risks.

Perhaps less well understood is the allocation of WHS responsibility in situations where AI is making, or materially influencing, safety-critical decisions. How does an officer verify that an AI-driven fatigue monitoring or hazard detection system is functioning as intended, when the system itself may not be able to be audited using traditional methods? To what extent can a duty holder rely on the expertise of third-party AI providers when trusting the AI systems they adopt in their operations? 

Safe Work Australia has explicitly identified the need for clearer definitions of WHS roles and responsibilities, particularly in cases where AI may contribute to safety breaches without direct human intent – suggesting that current regulations do not yet provide sufficient clarity. We anticipate that further regulatory guidance will be provided in this area and should be closely followed. 

Could you be required to implement AI to meet existing WHS duties?

It is no secret that AI use is increasing at a dramatic rate, with recent statistics showing 46% of Australian businesses are embracing new technologies, including AI. As AI continues to develop and its capacity to improve safety grows, the question all businesses should be considering is: 

Could I be required to implement AI in order to discharge my WHS obligations?

In certain contexts, the answer may be yes. The obligation to eliminate or minimise risk so far as is reasonably practicable requires consideration of the availability and suitability of ways to address the risk. As AI-based safety controls become more established and widely adopted, the bar for what is "reasonably practicable" will rise accordingly.

We are already seeing this shift reflected in regulatory guidance. For example, following a serious shuttle car incident at a coal mine, Resources Safety and Health Queensland issued a safety alert calling on Site Senior Executives to evaluate additional engineering controls, particularly AI or infrared camera systems. Safe Work Australia has stated that “Digital innovations offer significant potential to improve WHS outcomes through real-time risk detection and data-driven prevention strategies, with the possibility of eliminating some physical hazards entirely.” Given the safety benefits that AI-based controls can deliver, there may come a point where failing to implement such measures constitutes a breach of duty.

The regulatory and legislative response

The legislative landscape is moving quickly across multiple fronts, with recent developments including:

• The Federal Government has established the new Office of AI, which will design and legislate the new AI standards, and is pursuing AI safety in the workplace through the tripartite Artificial Intelligence Workplace and Employment Forum. 
• Safe Work Australia is undertaking a Best Practice Review of the model WHS laws with a report expected to be provided to WHS Ministers in August 2026. The review will consider AI and digital workplace transformation, with consultation revealing differing views among unions and industry on the need for additional safeguards. 
• In New South Wales, impending changes to the general WHS laws will impose a positive duty on employers to ensure that workers' health and safety is not put at risk by the allocation of work by a digital work system (see our related update here).  
• The Victorian Government has committed to introducing laws that protect workers from inappropriate and unfair uses of AI, including intrusive surveillance. This could include expansion of existing consultation obligations. 
• Safe Work Australia has released new guidance for duty holders in relation to artificial intelligence and digital technologies. 

These developments will need to be monitored closely as regulation attempts to keep pace with emerging technologies and AI adoption. The key will be ensuring that regulation strikes the right balance - protecting workers while also enabling businesses to test, adopt and refine new technologies to drive genuine safety improvements. 

What should duty holders be doing now?

• Treat AI systems as a WHS hazard: Identify and manage the risks under existing obligations.
• Monitor regulatory guidance: Safety alerts identifying AI-based controls should be tracked and considered carefully as they may begin to indicate the standard of care expected.

Take psychosocial risks seriously: Surveillance and AI tools are already capable of generating significant psychosocial hazards under existing law.

Please reach out to your usual HSF Kramer contact if you have any questions.