The Reality of Substitution
We recognize that machines are faster and more connected than human biology. This is not just about robots on assembly lines; it is about AI systems that can perform cognitive tasks, coordinate logistics, and generate content.
As tasks (not just jobs) are automated, the share of national income going to labour will fall, while the share going to capital (owners of the machines) will rise. Without intervention, this leads to extreme inequality and immiseration of large numbers of people.
“Imagine boat after boat arriving, each carrying thousands of brilliant workers willing to work for almost nothing. They take local jobs but hardly consume anything, no food or lodgings, and much of the pay they earn is shipped back overseas. Behind them come more boats carrying smarter workers, and behind them come even smarter workers again, but these are bringing robots.” — jnt discussions
The Great Displacement
Projected workforce demand across — Australian occupations, weighted by current employment. Based on AI-generated task erosion analysis using a Cumulative Task-Substitution model. Employment to the present is the historical record; the flattening into the turning point is already in the data. The projection begins at the dotted red line marked Now.
Move the pointer across the chart to see what drives the curve.
Explore how artificial intelligence may impact your career and what you can do about it.
Our Solution: The Human Utility Subsidy
Taxing the Machine Economy
We propose a transaction tax on the machine economy (Internet of Things) specifically compute. The engine of productivity of automated capital. This revenue is not used for a "dole" or UBI, but to fund a Negative-Payroll-Tax which encourages enterprises to find useful work for employees. As the machine economy and the GDP grows larger, this tax rate would decrease.
Incentivizing Human Employment
We will use a Floor-and-Trade mechanism to create tax offsets for hiring humans. By making humans cheaper to hire than robots for specific roles, we utilize market forces to find and preserve employment wherever human connection is valuable or desireable. This is not just confined to care, education, community safety, and craftsmanship.
Why Retraining Isn't Enough
Traditional parties promise to "retrain" workers for "jobs of the future." This is a rear-view mirror strategy that started several years ago with the "learn to code" refrain that is looking decidedly outdated now. "Technological unemployment" means discovery of labour-saving means outruns our ability to find new uses for labour.
We are seeing two linked disruptions:
- Automation of Capital: From dumb machines to autonomous intelligent systems.
- Commodification of Labour: The shift from full-time jobs to "contingent" gig work, where tasks are specified so precisely they become training data for the AI that will replace the worker.
The Paradox: Automate Harder, Employ More
It surprises people that the party most worried about automation wants more of it. We do, in the right places. Australia should grow its economy with AI and robotics in the two industries where the land itself is the advantage: mining and agriculture.
We have far more land than people to work it, and a long coast around it. The outback has areas that will never support a town, but fly-in fly-out crews with automation-assisted mining are already profitable there, and will become more so as AI and robotics advance. Robot farms and big irrigation projects can open country that was never worth farming by traditional methods, and will still employ people. The more projects, the more employment. The same holds at sea: the shallow waters along our coast carry minerals and resources we barely touch, and robot aquaculture can farm fish and seaweed at a scale no crew could. Work of this kind displaces nobody, because very few people are there now, and those who are gain employment opportunities they currently lack.
The same machines let us add more value before export. We already do it in places: most of our bauxite leaves as refined alumina rather than raw rock, and Whyalla and Port Kembla still make steel. But the pattern is thin. Of every hundred tonnes of iron ore we ship, less than one becomes Australian metal. The block was never just cheaper labour elsewhere; smelting and refining run on energy, and ours grew expensive. Robots close the labour gap, and cheap clean power closes the energy gap, so processing can come back to where the minerals and the power already are. Not every product pays its way: flour rarely earns its freight, which is why wheat leaves as grain. But our barley already leaves as malt for the world's brewers, and iron shipped as metal rather than ore would multiply the value of every tonne.
This does three things at once. It grows GDP, and with it the machine-economy tax that funds the Human Utility Subsidy. It creates more of the jobs machines still need people for — supervision, maintenance, transport, fixing what goes wrong — which is exactly where the chart above shows employment moving. And it extends a lead we already hold: Australian mining automation is among the best in the world.
There is a security aspect to this as well. A continent that looks empty and underused, held by a few million people with needs smaller than their neighbours', invites others to ask whether they could use it better. A continent in full productive use answers that question before it is asked. As global warming changes weather patterns and impacts the world's food supply, a nation that feeds more than itself is a nation the world wants left intact. By ensuring productivity we assist security.
The paradox comes apart cleanly. Displacement is machines taking work people already do, whereas expansion deploys machines at scale to do work no person can. We oppose the first when it is left unmanaged. We propose the second at national scale.