Don’t Waste the Downturn: Build the Skills AI won’t easily replace

New Zealand’s weak labour market is an opportunity to move more people into skilled, productive careers that will remain essential as AI reshapes other parts of the workforce. 

The opportunity in front of us 

New Zealand’s unemployment rate has reached 5.6%, with 171,000 people unemployed and 440,000 underutilised. The pressure is particularly visible among younger people: the NEET rate – the share of 15–24-year-olds who are not in employment, education or training – stands at 13.8%. 

Annual wage inflation is only 2.0%, while underutilisation has risen. Yet demand is beginning to return in parts of the productive economy: online job advertisements grew 7% in the year to June 2026, led by manufacturing, construction and primary industries. 

This creates a narrow but valuable window. We have available labour now, while the energy, infrastructure, engineering and industrial sectors are still building towards their next cycle of demand. Instead of waiting for shortages to return, we should use this period to connect people to paid, employer-linked training and recognised qualifications. 

These are the careers we should be selling 

The case is not only about filling future vacancies. We, as a nation, should be guiding more people towards careers that are comparatively resilient to AI and economically important to New Zealand. 

Reserve Bank research found that professional, managerial and administrative occupations carry relatively high AI exposure, while around 30% of workers face high combined exposure to AI and robotics. No occupation is immune to technology, but many skilled trades and technical roles bring together physical work, regulated competence, safety responsibility, site-specific judgement, fault diagnosis and human accountability. Those capabilities are much harder to replace with software alone. 

AI will change these jobs rather than remove the need for them. Electricians, line mechanics, engineering technicians, mechanical fitters and infrastructure specialists will increasingly use AI-assisted planning, diagnostics, design tools, automation and predictive maintenance. Their core trade or technical competence remains essential, while their technology capability becomes broader. 

That makes these sectors more compelling for young people and career changers: they offer practical work, nationally recognised skills, clearer progression and roles that should continue to matter in an AI-enabled economy for decades to come. 

The shortage has not gone away 

More than 100,000 people already work across New Zealand’s infrastructure system, and the national pipeline contains about $275 billion of planned projects. Engineering New Zealand estimates the country is short around 2,300 new engineers each year. Energy-sector planning also points to sustained demand across electricity generation, networks, maintenance and industrial operations. 

The recurring roles are not abstract: Capabilities like the examples below take years to build and are required across multiple sectors.  

• electricians, line mechanics, electrical and engineering technicians

• mechanical fitters, fabricators, welders, machinists

• project managers, surveyors, civil and mechanical engineers

• specialists in mechatronics, instrumentation and industrial maintenance. 

The recent downturn can actually deepen the future shortage. When project work slows, firms release experienced people, reduce graduate hiring and take on fewer apprentices. By the time investment returns, the pipeline is weaker and New Zealand is again competing for skills it did not develop. 

The training pipeline is moving backwards 

Workplace-based learner numbers fell 12% to 108,315 in 2025. Apprentice participation and new entrants also declined, while only 40% completed a qualification within five years. 

The reason is straightforward. Apprentices need real work, supervision and cash flow. When activity falls, employers cut training because learners initially consume time and capacity.  

Our system therefore trains fewer people when labour is available, then scrambles for capability when the market improves. 

Current support helps, but it does not solve project-cycle risk, limited supervisory capacity or continuity of employment. Nor should success be measured by how many people start. The real measures should be employee retention, qualification completion, progression into skilled work, and ultimately, wage growth. 

What needs to change 

We do not need another stand-alone training scheme. We need a practical system that links labour-market slack to industries with verified long-term demand: 

• Expand paid training during downturns. Increase apprenticeship, traineeship and technical-cadet places when youth unemployment and underutilisation rise, but only in occupations supported by credible project, retirement or vacancy demand. 

• Share the employment risk. Use licensed sector-based apprenticeship employers that can rotate learners across approved businesses and projects when one company’s workload falls. 

• Build skills into procurement. Major public programmes should publish workforce plans and assess bidders on realistic apprenticeship, graduate, technician and trainer commitments. 

• Fund completion, not just starts. Tie more support to first-year retention, competency milestones, recognised qualifications and progression into better-paid skilled work. 

• Make AI capability part of the trade. Combine enduring occupational competence with data literacy, automation, AI-assisted tools, cybersecurity, communication and judgement. 

Slack is an asset with an expiry date 

Not every unemployed person is suited to a trade or technical career. Aptitude, interest, health and occupational fit matter.  

But New Zealand has a clear opportunity to give more people a credible route into sectors that combine practical work, recognised qualifications, future demand and stronger resilience to AI-driven change. 

We ultimately have a choice to make: allow the downturn to weaken the future workforce, or use it to build scarce capability before demand accelerates. If we wait until the shortages are obvious, we will already be too late. 

Research and sources: The claims in this article were checked against current labour-market data from Stats NZ and MBIE; AI and robotics exposure analysis from the Reserve Bank of New Zealand; infrastructure workforce and project-pipeline evidence from Te Waihanga; engineering workforce analysis from Engineering New Zealand; and workplace training and apprenticeship data from Education Counts. The article also drew on sector workforce planning and vocational pathway research to test the direction of the argument, rather than relying on any single forecast.