In brief
It is clear from our recent work (and trite to say) that the pace of AI-related change at work is increasing across a range of industries – from tech, to financial services, to manufacturing.
AI presents opportunities for employers and employees alike, but the risk of AI-related change must be carefully managed to avoid legal and reputational exposure. While the "AI specific" regulatory landscape remains nascent in Australia, there are signs – including out of the Prime Minister's "AI in Australia's Interests" address at the University of Sydney this week – that increased regulation is on the horizon.
In any event, employers must understand that existing obligations may be triggered when implementing AI-driven change.
Key takeaways for employers
- Federal employment law does not deal directly with the topic of AI… yet. Following the Prime Minister's "AI in Australia's Interests" address, there are signs that may change.
- However, implementing AI-related change, including redundancies, will trigger existing obligations in many cases.
- Replacing jobs with AI may justify redundancy, but employers should ensure their rationale is tested and well documented.
- Consultation in relation to workplace change is essential to comply with award and work health and safety obligations.
- Employers should undertake an assessment of psychosocial health and safety risks arising from AI-related change. Safety regulators have moved from an era of education to enforcement.
- Redeployment obligations are onerous, particularly for larger organisations, and the Fair Work Commission can and will scrutinise employers' efforts.
In more detail
The regulation debate
Rumblings about the impact of AI on work and jobs have been getting louder over the past 12 months. While businesses have been quick to embrace the opportunities AI presents, unions have warned of mass redundancies unless workers are provided with job security and retraining. The Australian Council of Trade Unions has argued for a four-day working week so that workers can benefit from productivity gains born of technological advances;1 and in its submission to a Parliamentary inquiry into the adequacy of the National Employment Standards, the Australian Services Union called for a six month increase to paid notice for workers dismissed as a consequence of AI introduction.2
In April, Minister for Employment and Workplace Relations, Amanda Rishworth, made clear that the Government intends to take a tripartite approach to AI regulation, with input from government, employers and unions. She also played down the idea that unions might have a right to veto AI-related change.3
Most recently, on 15 July, the Prime Minister delivered his "AI in Australia's Interests" address at the University of Sydney. While the address did not provide detail on measures that may be implemented to govern workplaces in particular, it is clear the Government's priority is to manage AI-related change – not seek to prevent it – with a focus on job creation, not replacement. It appears the Government will seek to address this by way of regulation, with the Prime Minister expressing an aim of legislation early next year.4 Some commentators have noted that the speech is a sign the Government is prepared to use industrial laws to help "augment the workforce with AI rather than upend it".5 Watch this space.
Can the adoption of AI be used to justify redundancies?
The existing framework under the Fair Work Act means a dismissal will only constitute a genuine redundancy if:
- The employer no longer requires the job to be performed by anyone because of changes in the operational requirements of the enterprise
- The employer has complied with any applicable consultation obligations in a modern award or enterprise agreement
- Redeployment within the employer's business or that of an associated entity would not have been reasonable.
In the context of AI adoption, employers should be careful to distinguish between eliminating a role and changing how work is done. The introduction of AI may justify a redundancy where technology has genuinely impacted on the need for a particular position. However, if AI merely changes the way an employee's role can be best performed, a redundancy may be open to attack.
Employers should maintain evidence supporting operational change, including business cases, organisational charts, and Board or management minutes.
Consultation requirements and psychosocial risk
If a modern award or enterprise agreement applies, a dismissal will not qualify as a genuine redundancy6 if the employer has not complied with the consultation requirements contained in the relevant instrument.
Typically, consultation clauses require employers to notify impacted employees of definite decisions to make major changes in production, program, organisation, structure, or technology if those changes are likely to have significant effects on employees. Significant effects include termination of employment, but importantly go beyond that, and also include the need for retraining or restructuring. Employers should not assume that consultation obligations are only triggered at the pointy end of an AI-related dismissal. They may be triggered much earlier, at the point of adoption.
Regardless of whether an award or enterprise agreement applies, the force of work health and safety regulations dealing with psychosocial risk is now such that we recommend consultation, with additional support measures, in almost all cases.
Organisational justice and change management are on work health and safety regulators' hit lists.7 With psychosocial hazards firmly on the agenda for several years now, the approach of regulators has shifted from education to enforcement. While not AI-related, no case made this clearer than SafeWork NSW's issuing of a prohibition notice to the University of Technology Sydney as it sought to undertake a restructure in September last year. The notice, issued after an investigation prompted by an anonymous complaint, forced the University to pause its process until it could rectify alleged procedural and risk assessment-related deficiencies.8 One shudders at the thought of legal fees spent!
With this in mind, it is essential for employers to undertake psychosocial risk assessments in cases of AI-related change early and consider what measures should be put in place to manage identified risks.
Redeployment
Before finalising a redundancy, employers must consider whether it would have been reasonable to redeploy the impacted employee elsewhere within the employer's enterprise, or that of any associated entities.
In August last year, the High Court handed down its judgment in Helensburgh Coal Pty Ltd v Bartley & Ors,9 holding that the Fair Work Commission can make broad inquiries regarding redeployment. In that case, the inquiry was whether employees could have been placed into roles held by contractors. The underlying reasoning was that the way in which work is deployed may be up for question. This may be pertinent if AI-related change results in the "absorption" of certain responsibilities by technology at a "human cost", or where technology displaces workers performing certain functions but creates new opportunities in AI-related roles.
For larger organisations, the redeployment obligation can be particularly onerous because it extends to associated entities. Overseas opportunities may also be in scope (though tests of reasonableness apply).
Employers should conduct thorough searches and document the steps taken, as well as how workforce composition fits into the overarching workplace strategy.
2 Submission 108 - Australian Services Union.pdf
4 AI in Australia's interests | Prime Minister of Australia
6 For unfair dismissal purposes.
7 Psychosocial hazards | SafeWork NSW
9 [2025] HCA 29.