Taxing robots: the good news and risks about AI and government revenue - e61 INSTITUTE

TEXT START: AI may deliver a large economic dividend and disrupt the labour market, and the Prime Minister rightly had these in focus in his recent address on AI in Australia’s interests.

The Dissection

This is a fiscal-containment memo disguised as an AI analysis. It asks whether Australian government revenue survives a shift from wages to profits and from domestic labour to imported AI services. Its answer is cautiously affirmative: corporate tax may replace lost wage-tax revenue, productivity may enlarge the base, and GST may hedge against digitalisation.

The text never performs the central autopsy. It measures whether the state can keep collecting money, not whether mass productive participation survives. The labour market is treated as a tax input rather than the load-bearing structure of the post-WWII economy.

The Core Fallacy

It confuses revenue resilience with economic-system resilience.

Under the Discontinuity Thesis, tax receipts can remain stable or rise while AI makes most human labour economically unnecessary. Corporate tax captures profits; GST captures consumption. Neither restores the severed wage → employment → consumption circuit, nor gives displaced people ownership or control of the productive machinery.

The claim that imported AI services free domestic labour and capital for other uses also assumes that those other uses will require human workers at sufficient scale. That is precisely what P1, P2, and P3 deny. If AI dominates cognitive work and institutions cannot preserve human-only domains, the supposedly freed labour is not latent productive capacity. It is surplus.

Hidden Assumptions

• Wage losses become taxable domestic profits rather than offshore rents, shifted intellectual-property income, or lightly captured platform income.
• Displaced workers move into valuable new work instead of losing economic necessity.
• Productivity gains expand output and demand without severe concentration or demand failure.
• The tax system remains enforceable as AI firms globalise and intangible assets become harder to locate.
• GST can replace income-tax revenue without intensifying inequality or requiring compensating transfers.
• Foreign AI providers are taxed adequately at origin, despite extracting market, data, and user-base rents from Australia.
• Federal revenue is an adequate proxy for national stability, ignoring welfare costs, state payroll-tax losses, legitimacy, and political fracture.

The article acknowledges some of these limitations, especially its static modelling. That admission does not solve the problem. The excluded effects are where discontinuity occurs.

Social Function

Partial truth serving transition management and elite self-exoneration.

The article is not pure copium. It correctly identifies the distinction between a tax-base collapse and a dispute over which country receives taxing rights. It also identifies GST as more resilient to cross-border digital consumption than income tax.

But its practical function is to reassure fiscal managers that the state may survive AI even if workers do not. It converts mass displacement into a tax-design problem and treats the continued solvency of government as evidence of economic continuity. That is a category error with a spreadsheet attached.

The Verdict

A technically useful tax memo with a strategically fatal blind spot. It establishes only that AI may not bankrupt Australia’s treasury—and may even enlarge it—under its assumptions. It does not refute the Discontinuity Thesis.

If AI severs the wage-consumption circuit, GST becomes a meter attached to consumption increasingly sustained by transfers, while corporate tax becomes a toll on concentrated AI ownership. The state may learn to tax the carcass. That is not survival of post-WWII capitalism. It is fiscal adaptation to the disappearance of mass productive participation.