These workers say their jobs were replaced by AI. Could it happen to you? - SBS

URL SCAN: These workers say their jobs were replaced by AI. Could it happen to you? - SBS
FIRST LINE: Evidence shows AI is changing jobs, rather than completely replacing them. But the impact on the Australian workforce is likely to be unequal.

The Dissection

This is a transition-management article disguised as a risk assessment. It takes one worker’s restructuring story, an academic caveat, and a “keep humans at the centre” prescription, then frames displacement as workflow adaptation. Natalie’s experience suggests AI may have contributed, but the article does not establish that causally. Its central function is to acknowledge worker pain without confronting the competitive endpoint: firms use productivity gains to reduce labor requirements, not to preserve employment out of sentiment.

The Core Fallacy

“AI is changing jobs rather than replacing them” confuses the temporary form of displacement with its structural outcome. AI does not need to erase entire occupations instantly. It only needs to automate enough tasks that fewer people can produce the same output, after which restructures, hiring freezes, thinner teams, and weaker bargaining power do the rest.

The absence of wholesale replacement evidence is a lag observation, not a refutation of the Discontinuity Thesis. “Keeping humans at the centre” is moral language. Competitive systems centre cost, speed, scale, and control. Under P1, task compression is the entry point. P2 prevents stable human-only domains from surviving at scale. P3 arrives first as declining labor demand and only later becomes visible as mass job loss.

Hidden Assumptions

• Productivity gains will create enough new work to offset labor displacement.
• Employers will share efficiency gains with workers rather than capture them.
• Jobs remain intact when enough constituent tasks are automated.
• Government, businesses, and workers possess equal capacity to manage the transition.
• Reskilling can outrun the speed and breadth of AI deployment.
• Unequal impact remains a distributional problem rather than a participation-collapse problem.
• Today’s limited evidence represents a stable ceiling rather than an early phase.
• Employment can remain socially and economically necessary after production no longer requires most workers.

Social Function

Partial truth functioning as ideological anesthetic and transition management. The article accurately records an early lag phase: tasks are being reorganized, productivity is rising, and some workers are being displaced through restructures whose AI component is difficult to prove. But it turns a structural threat into a governable workplace adjustment, preserving the fiction that careful management can keep human labor central indefinitely.

The Verdict

The article is accurate about the present and misleading about the trajectory. It documents replacement before openly naming replacement: fewer workers required, workflows compressed, causality blurred, and losses distributed unevenly. The mass-employment circuit does not need a dramatic cliff. It can die through attrition—fewer openings, thinner teams, weaker wages, and recurring restructures. This is an early warning wrapped in reassurance, not evidence that the system is safe.