Artificial Intelligence has become surrounded by competing narratives. One side often predicts the wholesale replacement of workers, while the other side dismisses AI as an overhyped technology that will never live up to expectations.
The International Monetary Fund recently delivered what appears to be encouraging news for New Zealand. It concluded that New Zealand is among the countries best placed to benefit from artificial intelligence because of its skilled workforce and capacity to adopt new technology.
That optimism came with a warning though. While AI has the potential to improve productivity and economic growth, a significant proportion of New Zealand jobs may be affected by artificial intelligence, with many workers needing to adapt as technology changes the way work is performed. The IMF estimates that around one third of workers may be effected and it is likely to be particularly acute amongst women and younger workers.
AI is already transforming workplaces. It can analyse vast quantities of information in seconds, review documents, draft reports, automate repetitive administrative work and assist decision-makers in ways that would have seemed impossible only a few years ago. For a country like New Zealand plagued by low productivity AI does offer genuine opportunities to do more with the resources we have. But increased productivity should not be confused with simply reducing head counts in the workplace.
Some of the world's largest employers are discovering that replacing experienced workers with artificial intelligence is proving considerably more difficult than they planned for.
The Ford Motor Company invested heavily in AI-assisted engineering and manufacturing systems but it found that automation lacked something its most experienced engineers possessed in abundance - decades of accumulated practical knowledge, intuition and judgment. Recently Charles Poon, Ford's Vice-President of Vehicle Hardware Engineering, candidly explained "Mistakenly, we thought that by just introducing artificial intelligence ... that would produce a high-quality product." Ford has since brought back more than 300 experienced engineers (internally known as "greybeards") to mentor younger employees and to improve AI systems. Since doing so, Ford has reportedly reduced warranty costs, improved vehicle quality and achieved its highest industry quality ranking for more than a decade.
IBM has reached similar conclusions. While artificial intelligence has successfully automated routine human resources and administrative functions, it has confirmed that it continues to hire software engineers, sales professionals and client-facing staff because automation created demand for different human skills rather than eliminating the need for workers altogether. It has so far concluded that AI proves effective at routine tasks but it has proved much less effective where creativity, judgment and human relationships remain essential.
For 2026 IBM has planned to triple its entry-level hires in the United States. It’s Chief Human Resources Officer, Nickle LaMoreaux, said that investing in entry-level talent is essential to IBM’s long-term agility “if we don’t continue to invest in entry-level hires, what happens in 3-5 years”, and commenting further “there’s no pipeline, the well simply dries up”.
Closer to home, the Commonwealth Bank of Australia has also found that customer service could not simply be handed over to AI. While chatbots dealt efficiently with straightforward enquiries, many interactions ultimately required escalation to experienced employees. Last year CBA laid off 45 customer service staff and replaced them with an AI "voicebot” but it soon found that the AI system was unable to cope, which led to an increase in calls and it led to CBA to reversing the job cuts.
It is hard to object if AI removes repetitive administration, reduces waiting times and allows public servants to concentrate on work requiring higher levels of skill and judgment. But work is seldom a collection of tasks. It often requires exercising judgment, discretion, initiative and accountability. It often is assisted by “institutional knowledge”, either gained by experience or recognising whom to turn to. Those qualities are often impossible to reduce to a written process, let alone an algorithm. Responsibility still rests with a human decision-maker that can be held to account – not with an algorithm.
In this year's Budget, the Government signalled that AI would play an increasing role in cost saving and improving efficiency across the public service while continuing to reduce staffing numbers. It plans to further slash public service jobs by about 14 per cent (around 8,700 jobs) over the next three years, with a planned cost saving of about $2.4 billion.
Unfortunately, this government takes a pretty blunt approach to major ticket items; tiny tax cuts in 2024 which cost billions, tax cuts for landlords, again with billions in costs. We were told to tighten our belts, the government had nothing to meaningfully assist those hardest hit in the continuing cost of living crisis. Cost savings have taken similar blunt approaches; stripping pay equity from thousands and thousands of largely underpaid female workers. Other large investments in the future of New Zealanders faced similar approaches; slashing the quality of school lunches and scrapping large electric ferries for second hand Toyota type fuel guzzling ferries spring to mind.
If huge companies such as Ford, IBM and the Commonwealth Bank have rediscovered the value of experienced employees after attempting greater automation, the government and its agencies should be careful not to repeat the same experiment at the public expense. Read more...