Media release
Gone for Good: AI and the future of work
The AI revolution isn’t coming—it’s here, and it’s eliminating jobs. Unlike past automation, which replaced manual labour, AI is automating complex tasks in law, finance, journalism, and logistics, putting even high-skilled professions at risk.
Maxim Institute’s latest paper, Gone for Good: AI and the Future of Work, reveals the scale of this disruption and what it means for New Zealand. Conservative estimates suggest at least 10% of jobs could disappear in the short term, with the IMF warning AI will impact 40% of global employment—60% in advanced economies like ours. Secretarial, administrative, and customer service roles face the steepest decline, with up to 60% of tasks in these fields at risk.
Even high-paying white-collar jobs aren’t safe. AI is replacing junior legal researchers, automating financial analysis, and streamlining banking operations—JPMorgan’s AI tool COIN now does 360,000 hours of legal work per year. In retail and banking, AI-driven chatbots are handling customer service faster and cheaper than human staff. Manufacturing and transport industries are next, with warehouse robots and self-driving trucks set to upend logistics and supply chains.
This shift brings massive risks—rising unemployment, economic instability, and a growing divide as vulnerable workers, particularly women and ethnic minorities, bear the brunt of automation. But if harnessed wisely, AI could free people to focus on creativity, strategy, and innovation, leading to more fulfilling careers.
New Zealand must act now. Businesses must adapt, and education must evolve, or we risk being shaped by AI instead of shaping our future.
ENDS
This suffers from three fallacies at least.
First, current AI, the large language model form, is a lossy information summariser. Yes, it increases the productivity of basic legal research, and may appear to replace human researchers in the short term, but whether or not jobs are lost in the medium term depends term on the price elasticity of demand for legal research. I note there have already been several cases where lawyers have been embarrassed and censured for relying on AI, which often makes up legal precedents out of whole cloth. The fashion for dispensing with junior staff may yet die out.
A few jobs that are basically the lossy summarisation of information: journalism, legal research, helpdesk--may be at risk, but no job that requires correctness or precision is at stake. Nor is any job that requires judgement, not any job that deals with physical goods.
Second, it conflates AI and robotics. Self driving vehicles have been arriving "next year" since 2012. And truck drivers do far more than just drive trucks, just as accountants do much more than write journal entries and calculate depreciation. Warehouse robots are still very much at the single-robot trial stage, and they have been at that stage for at least three years. The fact that there have been no mainstream media articles talking about major rollouts of warehouse robotics suggests that nearly all of these trials have quietly failed.
Robotics is a whole bunch of difficult engineering problems that have to be solved individually--not least of them, safety in all situations--and then traded off against each other. Virtually none of the problems still to be fully solved in robotics have to do with AI.
Third, it suffers from out of date ideas about women. Women have the majority of tertiary qualifications these days. They are not a vulnerable minority. Also, ethnic minorities tend to work in occupations that are not at risk.