Sixty years ago, 'metal boxes' threatened to take over jobs. Now AI is | RNZ

TEXT START: In March 1968, in a rich, British accent typical of Australian news reporters of the time, Four Corners journalist Jim Downes asked the country: "When this circuit learns your job, what are you going to do?"

The Dissection

The article is a historical analogy functioning as narrative containment. It places AI inside the familiar story of computerization: frightening forecasts, institutional resistance, delayed adoption, retraining, and eventual creation of new work.

Its central move is to convert a question of ownership and productive survival into a question of timing. By emphasizing that earlier predictions arrived late, it presents institutional friction as “heartening” evidence. The article acknowledges disruption but leaves the post-WWII wage-to-consumption circuit largely unquestioned.

The Core Fallacy

The text mistakes deployment lag for systemic resilience.

That earlier automation took longer than predicted does not show that AI will preserve mass human employment. It shows only that physical, legal, and institutional defenses can delay technical displacement. Under the Discontinuity Thesis, those defenses are lag mechanisms, not reversals.

The article also assumes that displaced workers will be absorbed into “new things.” That assumption was more plausible when machines primarily attacked routine physical and clerical tasks. AI targets cognitive and knowledge work—the remaining foundation of middle-class wage participation. Retraining does not create scarcity if the retrained workers are competing with systems that can perform the same work more cheaply and at greater scale.

A strong job market today, and the failure of earlier three-year forecasts, are timing observations. They are not evidence against P1, P2, or P3.

Hidden Assumptions

• AI will generate enough new human labor demand to replace the work it eliminates.
• Human institutions can preserve stable human-only economic domains at scale.
• Retraining converts workers into complements rather than competitors of AI.
• Productivity gains will continue to flow through wages and broad consumption.
• Ownership and control of AI capital do not determine individual viability.
• A slower transition gives the majority a route upward rather than extending their dependence until displacement becomes unavoidable.
• “New possibilities” will be economically necessary, labor-intensive, and inaccessible to machines.
• Policy inquiries and new laws will manage the transition rather than merely document it.

Social Function

Primary classification: lullaby and ideological anesthetic.

Secondary classifications: transition management, prestige signaling, and partial truth.

The article is not crude propaganda. Its historical evidence is useful: technology does encounter institutions, and forecasts often miss the timing. But it turns that limited truth into a tranquilizing conclusion. It encourages readers to expect delay, adaptation, and retraining while avoiding the harder question: who controls the machines when human labor is no longer necessary?

The piece gives institutions a respectable script for managing public anxiety. It reframes structural dispossession as a temporary skills problem and treats uncertainty about future jobs as if it were evidence that those jobs will exist.

The Verdict

This is an intelligent pressure-release valve, not a serious account of the discontinuity.

It correctly identifies the repetition of the debate while missing the potentially decisive change in what is being automated. Computerization’s delayed impact does not guarantee that AI will recreate the human employment base it destroys. Under DT logic, institutions can slow mechanical death and cushion consumption, but they cannot preserve mass productive participation once AI dominates cognitive work and coordination fails.

The article’s question—what will you do when AI learns your job?—has no general answer in the text. The viable answers are structural: become a Sovereign, become indispensable as a Servitor, exploit the transition as a Hyena or Vulture, or secure power through energy, logistics, and maintenance. For everyone else, retraining risks becoming a queue for obsolescence.