New Zealand’s opposition has entered the AI-and-jobs debate with a plan that is unusual for what it puts first. On 9 September 2026, Labour leader Chris Hipkins and science, technology and innovation spokesperson Reuben Davidson released the party’s AI Action Plan, and the sentence most likely to matter to working readers is not about models or data centres. It is a commitment to “work with businesses and unions to make sure productivity gains show up in wages and conditions.”
🔍 THE BOTTOM LINE: Labour’s AI plan treats AI as a wage-distribution question as much as a technology question. Whether that promise survives contact with government is unknowable from an opposition document — but the mechanism it names, a seat for unions and employers in a long-term workforce plan, is specific enough to be checked later.
What the Plan Actually Commits To
The full release and a MinterEllisonRuddWatts legal summary published on 10 September lay out the structure. Labour would establish an Office of AI inside the Department of the Prime Minister and Cabinet, staffed, the party says, using existing resources. Among its listed duties: coordinating AI policy across government, setting “proportionate standards for responsible AI use,” and — the clause with the clearest jobs hook — preparing a long-term workforce plan “with working people, employers, unions and industry.”
Around that sit the familiar election-season items: an online safety regulator, a copyright framework so creators “control and are paid for their work,” rules requiring data centres to secure their own renewable energy, an AI in Education team for schools, and support for small businesses adopting AI safely.
Hipkins framed the upside in explicitly workplace terms: AI can “cut hospital paperwork so doctors and nurses have more time with patients” and let “workers and small businesses be more productive.” Davidson named the risk side, listing “job losses” first among New Zealanders’ concerns in the party’s release.
Why the Wages Clause Matters More Than It Sounds
The promise that productivity gains reach “wages and conditions” runs directly against the pattern New Zealand workers have experienced through earlier technology waves: productivity up, labour share of income flat or falling. The Guardian’s analysis of the Stanford evidence found the AI-era jobs apocalypse has not arrived in employment numbers, but work has changed anyway — and the gains have not obviously landed in pay packets.
Two prior commitments make the wages clause more than a slogan, if it is delivered. The workforce plan would be co-designed with unions and employer bodies rather than written for them, and the Office of AI would carry a senior minister’s portfolio responsibility, giving the jobs question an owner inside government. A whole-of-government work programme would be published, which is the kind of paper trail that makes a promise auditable.
Scepticism is warranted, and the party’s own framing invites it. Labour’s release says New Zealanders worry about “job losses, harmful content, data centres driving up energy costs, or a lack of accountability” — an admission that the political argument is being contested on exactly this ground, with the 2026 election approaching.
What It Would Mean for New Zealand Workers
The concrete reading for workers: under a Labour government, the party says wage-sharing would become an explicit test of AI adoption policy, and retraining would be planned years ahead rather than improvised after layoffs. The pledge is also a marker the current Government will be measured against. As our earlier analysis found, the Government’s own AI strategy identified a skills gap two years ago that has not budged — only 24 percent of New Zealanders have had any AI training while 91 percent of workers already use the tools.
Labour’s plan does not name funding for reskilling programmes, an omission that matters given the scale of AI-adjacent hiring overseas where most new roles do not require a degree. Whether “existing resources” can carry a national workforce plan is the question the policy document does not answer.
There is also a timing tension the party has not resolved: an opposition plan cannot bind employers today, and a workforce plan “with working people” would take years to write. Workers deciding how to respond to AI in 2026 will do so under whatever policy is current, not the one promised for after the election.
What Both Sides Agree On
The quiet consensus is that the jobs debate has moved past whether AI will change work. The New York Fed’s September survey of firm behaviour found only 4 percent of AI-using firms had laid anyone off — most retrained workers instead. The disagreement is about distribution: who captures the gains, and who bears the disruption. Labour’s answer is that government should tip the scales toward wages; the Government’s position, per its own strategy, is that adoption itself is the priority.
❓ FAQ
What did Labour actually announce? An AI Action Plan released on 9 September 2026 by Chris Hipkins and Reuben Davidson. Its headline commitments: an Office of AI in the Department of the Prime Minister and Cabinet, an online safety regulator, a copyright framework for creators, data centre rules, and a long-term AI workforce plan built with workers, employers and unions.
Is this law yet? No. It is an opposition policy platform ahead of the 2026 election. Nothing binds the current Government, and implementation would depend on Labour forming a government.
What does it promise for wages? The release commits Labour to “work with businesses and unions to make sure productivity gains show up in wages and conditions.” The Office of AI would also prepare a long-term workforce plan with unions and employers at the table.
How does this compare to the Government’s approach? The Government’s AI strategy focuses on adoption and skills but has not committed to a union-negotiated workforce plan. Its skills gap has widened since the strategy was published, and Labour’s plan is partly an argument that the gap needs an owner.
Does the plan fund retraining? The release does not attach figures to reskilling. The workforce plan would be the vehicle, and its funding is unspecified.
🔍 THE BOTTOM LINE
The plan’s most career-relevant sentence is also its most testable: productivity gains that show up in wages and conditions. If Labour wins and delivers the Office of AI and workforce plan as described, New Zealand would be one of the few countries where the division of AI’s gains is an explicit, measured policy goal. If the phrase dissolves into consultation round after consultation round, workers will have learned the usual lesson about election-year promises. Either way, the promise is now on the record, dated and specific enough to be judged against.
📰 Sources
- NZ Labour Party, “Release: Labour will make AI work for New Zealanders” (9 September 2026)
- MinterEllisonRuddWatts, “Labour unveils AI Action Plan” (10 September 2026)
- Singularity.Kiwi archive: NZ AI strategy skills gap analysis; New York Fed AI workforce survey coverage
