Beyond the headcount: why AI redundancies demand a different approach

One theme dominated almost every panel and keynote speech at the recent 2026 AFR Workforce Summit in Sydney: the impact of artificial intelligence (AI) on employment.

From an announcement by the Minister for Employment and Workplace Relations Amanda Rishworth of a new tripartite forum to further develop AI implementation policy, to Chief People Officers sharing some hard lessons learned through recent restructuring activity, the message was clear: instances of AI-driven redundancies appear to be accelerating and employers who handle them poorly face serious legal, financial and reputational consequences.

The first quarter of 2026 has brought with it a first wave of AI-driven job cuts across Australia. WiseTech Global announced 2,000 redundancies as part of an AI-driven restructure, while Block, the owner of Afterpay, cut more than 4,000 roles. Atlassian soon followed with 1,600 redundancies worldwide, framed as a shift toward AI-first operations.

Sydney now ranks third globally for technology job losses in 2026, behind only Seattle and San Francisco. Jobs and Skills Australia Commissioner Barney Glover has warned that automation-related losses will exceed 600,000 positions, more than 4 per cent of the national workforce [1]. So, this is no longer a theoretical concern, but a workforce planning imperative.

As noted above, Minister Rishworth has established the AI Employment and Workplaces Forum, a tripartite body bringing together government, unions and business groups to further develop policy on implementation of AI in the workplace, considering five themes: trust, capability, transparency, safety and productivity.

In parallel, the National AI Plan released in late 2025 commits the federal government to ensuring workplace rights remain fit for purpose, supporting digital skills uplift and giving workers a meaningful voice in AI adoption. NSW has moved further with its Work Health and Safety Amendment (Digital Work Systems) Act 2026 (NSW), making it explicit that AI systems can create workplace risks that employers must actively manage.

Unions have been vocal in their criticism of the government’s pace on AI regulation. The Australian Services Union (ASU), with members in administrative and clerical roles, has argued that its members are already experiencing the negative consequences of AI deployment at work, performing functions such as minute-taking, rostering and transcribing that are now being overtaken by AI.

In its submission to the (ongoing) parliamentary inquiry into the National Employment Standards, the ASU called for a minimum six months’ paid notice for workers who lose their jobs due to AI. Other unions, supportive of the proposition, also suggested proposals for a four-day working week and five weeks’ annual leave to share the gains of increased productivity [2].

The Finance Sector Union (FSU) has gone further, accusing the government of being “asleep at the wheel" [2] and warning that companies are outsourcing work to obscure the relationship between AI and job losses. ACTU secretary Sally McManus has called for an AI regulator “with teeth" [3].

Business groups, however, have pushed back.

Business Council of Australia chief executive Bran Black has cautioned against moving too fast, pointing to the European Union and Canada as examples of jurisdictions that regulated early and are now rolling back their positions after losing investment [4].

The tension between these positions will shape the trajectory of Australia’s regulatory response in the months ahead.

A significant risk for employers undertaking restructures (including those which are AI-driven) lies in psychosocial safety.

In September 2025, SafeWork NSW issued a prohibition notice halting a restructure at the University of Technology Sydney (UTS) on psychosocial risk grounds. UTS had been undertaking significant role reductions as part of a broader cost reduction program when SafeWork identified a number of concerns with the approach adopted by UTS. The estimated cost while the prohibition notice remained in force was $3 to $4 million per month.

With psychosocial hazard regulations now in force across all Australian jurisdictions, every employer undertaking workforce restructuring must actively manage work-related mental health risks. These obligations apply to businesses of all sizes.

Beyond redundancy processes, AI itself is also creating new psychosocial hazards.

Minister Rishworth has flagged that AI can accelerate tasks in ways that impose unsustainable workloads and that AI-powered surveillance increases psychological stress and the risk of psychosocial injury. Business and government both have a responsibility to monitor and manage these emerging risks.

Organisations with blended workforces face a further layer of complexity following the High Court’s decision in Helensburgh Coal Pty Ltd v Bartley [2025] HCA 29.

The Court held that the Fair Work Commission is permitted to inquire, when hearing an unfair dismissal, into whether an employer could have made changes to how it uses its workforce, including by terminating contractor arrangements, so as to create or make available a position for a redundant employee.

Critically, the concept of “redeployment” under s 389(2) of the Fair Work Act 2009 (Cth) does not require a vacant position to exist; it contemplates reorganisation or rearrangement of the employer’s enterprise.

In the AI context, this means that where an employer uses automation to consolidate tasks previously performed by a blended workforce of employees and contractors, simply pointing to the absence of a vacant role will not establish a genuine redundancy.

Employers must assess the attributes of the redundant employee, the composition and contractual arrangements of their workforce, and whether it would have been reasonable in all the circumstances to redeploy rather than terminate. Failure to undertake this analysis risks employment litigation exposure and potential WHS enforcement action if the psychosocial impacts on displaced workers are not properly managed.

For employers considering or undertaking AI-related workforce changes:

• Conduct psychosocial risk assessments before commencing any restructure, not during or after.
• Analyse work at the task level, not the role level. Early indications from recent high profile restructures suggest that identifying specific tasks suitable for automation yields better outcomes than eliminating entire roles.
• Consider further investment in retraining and redeployment pathways, not only to meet WHS obligations but more beneficially, to retain talent.
• Assess whether AI tools already in use may themselves be creating psychosocial hazards through algorithmic performance monitoring, excessive workload allocation or reduced worker autonomy.
• Document everything. From a WHS perspective, regulators expect evidence of practical steps taken to manage risks, from regular check-ins with employees to workload adjustments.
• For organisations with blended workforces, the High Court’s Helensburgh Coal decision confirms the FWC can examine whether contractor arrangements could have been reorganised to redeploy a redundant employee. Where AI is consolidating tasks across employee and contractor categories, employers must assess redeployment options before concluding a genuine redundancy exists.

The AI revolution is not slowing down. Employees and their unions are hypervigilant of any related impacts from the implementation of this technology.

Whilst the Minister has signalled the intention to conduct a gap analysis of workplace laws and applicable regulatory frameworks, it’s clear that the landscape will continue to evolve. Employers who act proactively, embracing AI’s potential while managing its human impact, will be best placed to navigate what comes next.

References:

[1] Link to report by Jobs & Skills Australia, September 2025:

jsa_gen_ai_capacity_study_final_analysis_papers_03092025.docx

[2] The Australian “Union demands four-day week and six months notice for AI job losses”.

[3] Australian Financial Review "White-collar unions clash with minister".

[4] Ibid.

[5] Ibid "AI job losses could now exceed 4pc".

The views expressed in this article are general in nature only and do not constitute legal advice. Please contact us if you require specific advice tailored to the needs of your organisation.

For more insights from the Kingston Reid team on the workplace law issues facing organisations in 2026, head to our Publications page to access our 2026 Workplace Insights report.

Photo by Growtika on Unsplash.