AI adoption among businesses in Singapore rises; digital economy accounts for larger share of GDP

TEXT START: AI adoption among Singapore businesses rose to 23.5% in 2025, with small and large enterprises increasing usage and the digital economy growing to $144.1 billion, contributing 19.3% of GDP.

The Dissection

This is a transition-management narrative dressed as a growth report. It treats rising digital output, higher tech wages, and temporary tech-employment growth as evidence that AI is strengthening the labor economy. The same text supplies the counterevidence: firms cite productivity and cost reduction, 64% of AI-using workers automate processes, six in ten save up to an hour a day, and agentic systems are being developed with minimal human supervision. The article measures the expansion of the machine layer, then quietly treats the humans servicing that layer as proof that humans remain economically necessary.

The digital economy grew from $136.5 billion to $144.1 billion, while tech employment grew 3.8% to 222,200. That is not a contradiction; it is early transition behavior. Output and implementation labor rise while substitution infrastructure thickens. Demand concentrated in non-tech sectors shows where AI is being operationalized, not that those jobs are permanently protected. The text also contains a temporal defect: a report framed around 2025 figures suddenly cites worker AI usage in 2026. That weakens precision but not the direction of the evidence.

The Core Fallacy

It conflates aggregate growth with broad human viability.

Under DT mechanics, AI can increase GDP, lower costs, raise the wages of scarce technical workers, and still destroy the mass employment-to-consumption circuit. Growth in digital value-add measures machine-mediated production, not proof that the majority retain productive participation. The 60% tech wage premium shows scarcity rents around AI capital, not a durable path for ordinary workers.

The second fallacy is treating displacement as a training deficit. Training can improve a worker’s position during the implementation bottleneck. It cannot guarantee human indispensability once AI becomes cheaper, faster, and more scalable across cognitive work. Upskilling 140,000 workers is transition management, not a solution to P1, P2, or P3.

Hidden Assumptions

• Productivity gains will flow into hiring, wages, or shorter work rather than labor elimination and margin capture.
• High demand for AI, data, and cybersecurity workers is permanent rather than part of a buildout and integration phase.
• Workers can continually reskill faster than AI expands the set of automatable tasks.
• Human-plus-AI labor remains economically superior to integrated AI agents once systems become reliable.
• Digital growth in non-tech sectors translates into durable human roles rather than temporary implementation, verification, sales, compliance, and exception-handling niches.
• Government training and employer support can overcome time, budget, and coordination constraints at national scale.
• GDP growth and median tech wages are adequate proxies for worker security, despite the article never addressing ownership, bargaining power, or distribution.
• Gradual adoption percentages indicate a manageable transition rather than a compounding diffusion curve.
• The majority can become either owners/controllers of AI capital or indispensable servitors. The data establish neither.

Social Function

Primary classification: transition management and ideological anesthetic, with partial truth and prestige signaling.

The article gives institutions a respectable script: Singapore is digitalizing, workers are more productive, wages are high, and the state is funding training. It converts a structural power shift into an administrative skills gap. That framing reassures firms and policymakers that the problem is insufficient preparation, not the erosion of labor’s necessity.

It is not fabricated. Adoption, digital output, employment, and wage gains can all be real. That is precisely why the narrative works. Its partial truth conceals the terminal implication: a growing machine economy can make human labor less central, not more.

The Verdict

This is an early-warning document disguised as good news. Singapore’s digital economy is expanding because AI is increasing efficiency and reducing the labor required per unit of output. The rise in tech jobs is a temporary implementation moat around a system whose strategic direction is labor substitution. Training programs may delay social death and create servitor niches; they cannot reverse the discontinuity. The report celebrates the machine’s rising GDP contribution while failing to ask who will still be needed when the machine no longer needs as many people.