AI cuts in Australia: signal or smokescreen? What HR needs to do now
More than 1,000 tech jobs were cut in Australia in recent months, with AI cited as the reason. Atlassian let go of 500 staff locally as part of 1,600 global redundancies. Block cut 4,000 worldwide, reportedly including 700 Australians. WiseTech shed 2,000, and Telstra axed 200 roles in its AI joint venture with Accenture.
Executives say the story isn't "AI replaces people," but a shift in skills and mix. Atlassian's CEO put it plainly: AI changes "the mix of skills we need or the number of roles required in certain areas." WiseTech's chief executive went further: "the era of manually writing code as a core act of engineering is over." That's unsettling for employees - and it puts HR on the front line.
Is this real transformation or AI-washing?
Some analysts see AI being used as cover for broader restructuring and investor pressure. One advisory firm's modelling suggests only about 1% of cuts stem directly from AI productivity gains. At the same time, companies that announced layoffs saw quick share price rebounds - Block up 20%, WiseTech up 11% - even as Atlassian fell. Markets reward "efficiency"; AI provides a convenient narrative.
There's also cost and capability building behind the scenes. AI can be expensive to implement and slow to pay off. Morningstar uses AI to automate menial data gathering for analysts, but isn't cutting junior headcount. Call centres, often seen as first to automate, are still steadily hiring humans.
The labour market signal HR can't ignore
Adoption is spreading: nearly one in three Australian firms are already using AI for advanced tasks like demand and inventory forecasting. Anxiety is keeping pace: almost a third of workers believe their job could disappear due to AI. Early signs show pressure on entry-level and "foundational" product roles, mirroring US trends in finance, computing, sales and admin.
Locally, recruiters report smaller consulting and marketing firms trimming junior intake and slotting AI into short projects. Finance graduates are shifting from analytics toward human-facing roles such as financial advice. NAB data indicates white-collar unemployment has started rising faster than blue-collar, even if the latter remains higher overall. The work hasn't vanished - the task mix has.
Hard lesson from the creative front
Voice actors like Teresa Lim are seeing a more direct threat. An AI dupe can be built from a 15-second clip of her voice - and there's no current Australian legislation clearly blocking that. For HR, this isn't theoretical. It's a live issue for consent, IP, and fair compensation when an employee's likeness, voice, or work output trains a model.
Your 90-day HR action plan
1) Map tasks, not titles
- Break critical roles into tasks. Flag what can be automated, augmented, or left human-only.
- Redesign job descriptions around "AI-augmented" work. Define decision rights, quality gates, and human-in-the-loop checkpoints.
- Set clear standards for prompt quality, model use, and documentation.
2) Build skills faster than you cut headcount
- Stand up short, role-based upskilling in data literacy, prompt design, workflow automation, and model oversight.
- Protect the early-career pipeline. Pair juniors with AI to expand throughput, not replace them. Use rotations and real project reps.
- Codify new career paths: practitioner → systems thinker → AI workflow lead.
3) Governance that's actually usable
- Write policy on likeness/voice/data use with explicit consent and opt-out. Include rules for vendor training data and model fine-tuning.
- Approve a vendor list. Require audits for bias, data handling, and output accuracy.
- Create an "AI exception" path: documented pilots with time-boxed reviews and kill-switches.
4) Workforce planning with scenarios, not slogans
- Model three cases: no AI, augmented (10-25% task shift), and automation-heavy (25-40% task shift). Plan redeployment before redundancies.
- Use attrition to right-size where possible. Prioritize internal moves to product ops, QA, AI workflow design, and client-facing roles.
- Align hiring bar with the new work: fewer solo builders, more integrators and reviewers.
5) Performance and rewards that fit AI-era work
- Shift goals from "hours spent" to cycle time, quality, risk, and customer outcomes.
- Reward reusable workflows, clean data, and documented prompts - the new leverage.
- Add competencies: model selection, validation, exception handling, and ethical judgment.
6) Change comms that earns trust
- Be specific: which tasks change, which roles are safe, and which are at risk over what timeline.
- Separate "efficiency gains" from "role eliminations." Don't let "AI took your job" become the scapegoat.
- Offer real support: internal gigs, scholarships, certifications, and manager-led coaching.
Metrics that matter
- Adoption: percent of priority workflows augmented by AI; number of active users per week.
- Throughput: cycle time per task; tickets or stories closed per FTE; SLA adherence.
- Quality: error/defect rates; rework; customer CSAT/NPS where AI is in the loop.
- People: internal mobility rate; junior-to-senior ratio; time-to-productivity for new hires.
- Risk: policy exceptions; data incidents; audit findings on bias and accuracy.
What to tell your executives
"AI is changing task mix faster than job count. We'll redesign roles around high-impact workflows, use attrition and redeployment first, and invest in skills that lift output and quality. Where roles end, we'll say so plainly and support transitions. Our goal is fewer handoffs, better decisions, and measurable gains - with guardrails that protect people and the brand."
Practical resources for HR
The bottom line
AI is starting to rewire work, but many "AI-driven" cuts are standard restructuring in new clothes. HR's job is to make the work visible, move people toward higher-value tasks, and put real guardrails in place. Do that, and you protect both performance and trust - no buzzwords required.
Your membership also unlocks:
