Australia has chosen its side in the global argument about how governments should manage AI’s arrival at work: persuasion first, law later. On Thursday, 12 November, Workplace Relations Minister Amanda Rishworth launched a draft Best Practice Guide on the Adoption of AI in the Workplace — eight principles under three themes (plan, enable, embed), built from roundtables with Microsoft, Commonwealth Bank, Telstra, Kmart, Blackmores and Ramsay Health Care, as The Mandarin reported from Canberra. Employers get a practical roadmap; workers get the promise that they will be consulted.
But the same document, buried deeper in the text, concedes the thing the launch event did not: entry-level work “may be the first to be automated,” a phrase The Guardian’s report on the launch elevated to its central finding. The guide warns the technology’s hazards include “intrusive worker surveillance, poor rostering practices, intensified or unrealistic performance standards” and discrimination — against younger, older and First Nations workers, women and people with disability.
The guide’s pitch and the market’s reality
Rishworth’s argument, as she put it in the government’s release: “The Best Practice Guide makes it clear that AI works best when the workforce is at the centre of the adoption process.” The department’s overview cites its own case studies — Brisbane Catholic Education reporting 71% of educators saving admin time, CBA running 32,000 employees through AI career tools, Blackmores compressing a three-day data-entry process into ten minutes. The guide is open for consultation until 27 January 2027.
The pitch is that AI can mean fewer hours and higher wages. The timing is awkward, because the Australian labour market is currently running the other experiment. In the space of a single month, NAB cut 2,500 roles, Canva announced 1,200 layoffs, and Qantas moved 850 corporate jobs into an Accenture outsourcing arrangement that ships much of the work to India first — all attached, in whole or part, to AI modernisation stories. The government’s own release frames the guide as the answer to workplace “transformation” that is already under way, with or without best practice.
At a parliamentary hearing into the big four banks this week, former industry minister Ed Husic put the question the guide cannot answer by consultation alone: “Who receives the benefits of higher productivity? Does it mean better services and lower prices for customers, secure jobs with higher wages for workers, or just lower labour costs and higher returns?” — as quoted in The Guardian’s coverage.
Greens leader David Shoebridge went further, arguing a worker who loses their job to an algorithm currently has “no legal recourse at all,” and that sharing AI’s benefits through pay and hours “needs to be written into law instead of left to the goodwill of employers.” The Reserve Bank, for its part, published its own monitoring paper this week finding “little evidence to date” of AI disrupting the labour market overall, with “tentative signs” of impact concentrated in clerical, programming and call-centre roles — which is to say, the entry and mid-level rungs the guide itself flags.
What the guide does — and what it doesn’t
Voluntary best-practice guidance is a real instrument when it shapes procurement and norms before deployment: the eight principles include cooperative planning and worker engagement, designing work and jobs, establishing AI governance, and proactively managing risks to work health and safety. Companies now have a document to point to when boards ask what “good” AI adoption looks like, and unions have a benchmark to hold them against.
But the guide’s release lands alongside a pattern Singularity.Kiwi has tracked across the Tasman all year: Australia’s wider AI workplace gap, where the think-tank and union critique in May was that regulation was repeating social media’s slowness; the Reserve Bank’s November finding that AI slows hiring rather than destroying jobs; and the transport minister’s warning to Qantas that AI-branded restructuring carries political cost. Consultation is the tool you use when the law hasn’t caught up. The people losing the jobs notice the difference between the two.
For New Zealand, the practical read is direct. This country has no AI-specific employment law either — the Employment Relations Act’s good-faith consultation duties, written for restructures of the 2000s, are doing the load-bearing work. The Australian guide’s most exportable line is not the eight principles but the admission tucked inside them: automation arrives at the bottom of the ladder first, precisely where the workers have the least bargaining power. A first rung that’s automated away doesn’t grow back, and no consultation framework in either country yet answers what replaces it.
Sources: The Guardian — Fewer hours and higher wages – or more burnout and surveillance: employers told to prepare for AI, The Mandarin — AI at work needs collaborative workplace transformation, government says, RBA — Monitoring the early effects of AI on the labour market
