This is a researched industry analysis. It distinguishes documented events from interpretation. It does not claim that every company, manager, or redundancy decision is unlawful.
AI Didn’t Break Australian IT. The Growth Model Did.
Artificial intelligence has become the preferred explanation for almost every unpleasant decision in technology.
A company cuts staff? AI.
A junior hiring programme disappears? AI.
A local delivery team is replaced by an offshore centre? AI-enabled efficiency.
A promotion is delayed, a salary budget is frozen, or a project team is reduced to its thinnest possible shape? The language changes, but the answer is often the same: the future has arrived, and somebody else has to pay for it.
The future is real. The explanation is often incomplete.
Australia’s Department of Employment and Workplace Relations published its first major study of AI and employment in July 2026. It found no broad AI-driven upheaval in the labour market. Jobs in occupations judged more exposed to generative AI had grown more slowly than jobs in less-exposed occupations, but the department described the evidence as an early, modest signal rather than proof of mass AI job loss. ABC reporting on the study noted that Australian software-development employment had increased by 25 percent since ChatGPT was introduced.
That should have ended the lazy story that AI has already erased Australian IT.
It didn’t, because the story is useful. “AI made us do it” sounds like physics. It hides the human choices between a new tool and a redundancy letter.
The technology is new. The incentive is old.
Public companies are built around growth. Revenue must rise, margins must hold, investors must be reassured, and the share price must tell a story about the next decade rather than the last quarter.
Growth itself isn’t a moral failure. A company that doesn’t improve its products, win customers, or manage costs can’t employ people forever. The problem begins when growth becomes an unlimited demand rather than a means to build a durable business.
Under that model, every efficiency becomes a headcount opportunity. If an engineer can produce more with an AI assistant, the company doesn’t have to choose between shorter hours, better quality, more training, and higher pay. It can choose the most financially extractive option: keep the target, remove the people, and call the remaining workload productivity.
That isn’t an inevitable consequence of AI. It’s a management decision.
The pattern is visible in the public record. In March 2026, Atlassian announced 1,600 global job cuts, including 480 roles in Australia, and said funds freed by the cuts would support investment in AI and enterprise sales. The same ABC report described the cuts as 10 percent of the global workforce. The Guardian reported that Block, WiseTech Global, and Atlassian all reduced staff while linking the future of their businesses to AI, efficiency, or a changing software market.
Those events don’t prove that AI performed the work of every person who lost a job. They show something more uncomfortable: corporate leaders can use AI as both a real technology and a financial narrative. It can change tasks, justify reorganisation, reassure investors, and make a workforce reduction sound like an act of technological realism.
The “unlimited growth” machine has a human input
The growth model depends on inputs that rarely appear in the headline announcement:
- Years of training that employees paid for with their time
- Institutional knowledge built through failed projects and difficult clients
- Junior workers who absorb routine tasks before they become senior specialists
- Managers who translate strategy into workable delivery
- Engineers who carry production risk after the announcement is over
- Families that absorb income shocks when a company changes direction
When a business removes people to fund AI, it treats those inputs as if they were disposable. The balance sheet records a saving. The community absorbs the cost through unemployment, lost confidence, delayed housing decisions, retraining bills, and people leaving a profession they once expected to build a life in.
That is why the argument cannot stop at “learn AI.” A worker can learn every new tool and still lose their job if the company’s objective is to deliver the same output with fewer people. Reskilling is useful, but it cannot solve a distribution problem by itself.
Australia is exposed to more than automation
Australian IT workers face a compound pressure:
- Automation: Software can reduce the time needed for routine analysis, support, documentation, coding, testing, and content production.
- Offshoring: Work can be moved to lower-cost labour markets while local staff retain client accountability and quality risk.
- Financial pressure: A company can reduce headcount because growth slowed, debt increased, an acquisition failed, or the share price fell, then present AI as the strategic answer.
- Weak career pipelines: Employers can stop hiring juniors while demanding that experienced workers become productive immediately.
- Work intensification: The remaining workforce is expected to supervise more automated output, correct more errors, and deliver more change without a matching reduction in targets.
The result is not simply fewer jobs. It is a worse bargain for the people who remain: more responsibility, less control, thinner teams, and a career ladder with missing rungs.
The corporate story is not the same as the worker story
A chief executive can say that AI will create better jobs in the long term. That may even be true for some workers. It doesn’t answer what happens to the person whose role disappears this month, or to the graduate who can’t get the entry-level experience that senior roles require.
A corporate strategy also uses time differently from a human life. Leadership can describe a three-year transition. An employee has rent due next week, a mortgage, a visa, dependants, health costs, and a professional identity that may have taken a decade to build.
The language of transformation becomes cruel when it treats those time scales as interchangeable.
What should change
The response cannot be a ban on every efficiency tool. It should be a demand that the gains are shared and the risks are governed.
Companies adopting AI should publish, at least internally:
- Which tasks and roles will change
- Whether the aim is augmentation, reduced hiring, redeployment, or redundancy
- What error rates and quality controls apply
- How performance data will be used
- What training happens during paid work time
- How affected workers can move into new roles
- What happens to workload when headcount falls
- Whether executive incentives include retention, sustainable workload, and redeployment outcomes
Workers should ask those questions before the tool is deployed, not after the redundancy process begins. The ACTU has called for employer transparency and consultation when workplaces introduce AI. That position is not radical. It is the minimum needed to stop a technical deployment becoming an invisible employment decision.
Governments should measure more than aggregate employment. They should track entry-level hiring, involuntary exits, occupational transitions, wage movement, job quality, local capability, and the distribution of productivity gains. A national labour market can look healthy while a profession’s junior pipeline collapses.
The point is not to reject the future
There is no return to an IT industry without automation. There was never an IT industry without automation.
The choice is about who controls the transition and who receives its benefits.
AI can remove pointless work, improve accessibility, reduce administrative load, and help small teams do valuable things. It can also become a convenient mechanism for transferring wealth and control from skilled workers to shareholders and executives.
Both statements can be true.
The task for Australian IT professionals is not to pretend the technology is harmless. It is to refuse the idea that every corporate decision made under its banner is inevitable. A tool has no right to determine the terms of a person’s life. People, workplaces, unions, regulators, and communities still decide those terms.
The first step toward reclaiming work is to name the decision correctly. When a company uses AI to cut costs, that is not “the market” acting like weather. It is a choice about who benefits from productivity and who carries the loss.
Sources and method
This article was prepared on August 8, 2026, using public material available at the time of writing:
- The AI and employment in Australia report — Department of Employment and Workplace Relations — government findings and limitations
- AI and employment in Australia — Department of Employment and Workplace Relations — report description and data framework
- Atlassian cuts 1,600 jobs from global workforce as AI slashes labour needs — ABC News — Atlassian’s announced cuts and stated investment rationale
- Will AI take Australian jobs, or is it just an excuse for corporate restructure? — The Guardian — reporting on Australian technology cuts and competing explanations
- AI: Unions put corporate Australia on notice — ACTU — union position on consultation and transparency
The article’s discussion of growth incentives and worker power is analysis, not a claim that every company follows the same strategy.
This is industry commentary, not financial, employment, or legal advice.
Related reading
- The AI Layoff Paradox: Australia Says “Not Yet” While Workers Lose Jobs
- The IT Career Ladder Is Being Pulled Up Behind Us
- Reclaiming Work: A Survival and Power Plan for Australian IT Professionals
⚠️ The Cost of Waiting
Australian MSP workers who negotiated using our salary data earned an average of $8,200 more per year. Every month you wait is ~$683 left on the table.
💰 Check if you're underpaid →🎁 Free Resource: Red Flag Checklist
12 contract clauses every Australian MSP worker should flag before signing. Includes non-compete traps, sham contracting indicators, and on-call gotchas.
Was this helpful?