A professional in New Zealand who built a career earning $120,000 a year is now making a third of that. Not because they made a mistake. Not because their industry collapsed. Because AI tools automated the work they were paid to do, and the market found no use for their skills at the price they had come to expect.
A Stuff investigation profiles workers on both ends of the disruption curve — those whose professional incomes have been slashed by AI displacement, and those just entering the workforce who are trying to figure out which skills will still be valued in five years. Together, the profiles capture an economic disruption that New Zealand's employment support systems were plainly not designed to handle.
The affected workers are not factory floor labourers — the people most employment disruption narratives have historically focused on. They are content creators, legal researchers, data analysts, and coders. People with tertiary qualifications and professional track records. People who did everything right by the conventional career playbook, only to find that the playbook had been rewritten overnight by systems that can now perform their core tasks in seconds for cents.
The speed of disruption is what makes this different from previous technological displacement. When manufacturing automation hit New Zealand's industrial sector in the 1980s and 1990s, the adjustment was painful but its contours were legible over years. Workers and governments could see what was coming. Retraining programmes, however imperfect, had time to be designed and deployed. The current AI disruption is moving in months, not years, and the breadth of occupations affected is expanding continuously.
New Zealand's social safety net — structured around Work and Income, retraining subsidies, and industry-specific redundancy arrangements — was designed for a labour market that moved at a different pace. The Jobseeker Support payment averages around NZ$350 per week for a single adult. A professional who was earning $120,000 annually faces a cliff of extraordinary steepness. Retraining pathways take months to years. The trades, healthcare, and education sectors currently least exposed to AI require extended study and accreditation that is simply not accessible to someone with a mortgage and immediate income needs.
The Luxon government has not signalled any comprehensive policy response to AI-driven labour displacement. The prevailing philosophy — that labour markets self-correct, that growth creates new jobs, that individual retraining is the appropriate answer — is being stress-tested by a technological shift whose pace it did not anticipate.
For Australia, the same dynamics are arriving with a slight lag. The same occupational profiles are being disrupted. The same gap between the speed of displacement and the capacity of social support systems to adapt is opening. When a Wellington law firm cuts its junior research staff after rolling out AI review tools, the effects are not aggregate statistics. They are individual households recalculating everything they planned.
The workers in the Stuff investigation are not asking for protection from change. They are asking for acknowledgment that the change is happening, and that the systems meant to support them through it are not remotely adequate for the task.
