AI not a job killer, say most Swiss businesses - SWI swissinfo.ch

TEXT START: Now widely adopted in day-to-day operations by the vast majority of small and medium-sized enterprises (SMEs), artificial intelligence (AI) is becoming an integral part of everyday business activities without, for the time being, raising fears of any major upheaval, particularly in terms of employment.

The Dissection

The article converts a snapshot of SME sentiment into an implied forecast of economic stability. Its figures show adoption, not harmlessness: AI use rose from 55% to 74%, expanded into process optimisation and data analysis, and is already producing time savings. The fact that only 12% currently anticipate job losses proves only that displacement has not yet become an explicit managerial expectation.

The text mistakes visible layoffs for the mechanism of destruction. AI can reduce future hiring, eliminate vacancies through attrition, compress hours, weaken wages, and allow the same firms to produce more with fewer people without announcing mass redundancies. Employment can remain temporarily stable while human productive necessity is being hollowed out.

The Core Fallacy

The central error is treating present equilibrium as structural permanence. Under the Discontinuity Thesis, AI does not need to trigger immediate unemployment to kill the post-WWII circuit. Once cognitive automation delivers durable cost and performance advantages, competition forces firms to convert efficiency gains into lower costs, higher output, reduced hiring, and eventually lower labor demand.

“Time savings without impact on employment” is therefore a lag condition, not a refutation. It describes the cushioning phase before competitive pressure makes redundant labor economically visible. A firm may preserve headcount today because contracts, demand, regulation, and organizational inertia delay adjustment. Those defenses slow the kill mechanism; they do not disable it.

Hidden Assumptions

• That current employment levels matter more than future hiring, wages, hours, and career ladders.
• That productivity gains will remain inside firms rather than being competed away through prices and margins.
• That SMEs can independently avoid restructuring while competitors adopt the same capabilities.
• That AI’s present task mix is its ceiling rather than an early deployment phase.
• That “no fundamental change” in the workplace means no loss of human economic necessity.
• That managerial expectations are a reliable measure of technological impact.
• That Swiss institutional stability can preserve human-only productive domains at scale.

These assumptions smuggle permanence into a survey designed to measure the present.

Social Function

Classification: partial truth, ideological anesthetic, and transition management.

The partial truth is real: AI adoption initially appears as augmentation, workflow assistance, and efficiency improvement. The anesthetic lies in presenting this transitional calm as evidence that employment is safe. The article gives institutions a reassuring narrative in which businesses can absorb automation without confronting the eventual collapse of mass productive participation.

Its most revealing phrase is “for the time being.” That qualifier quietly admits the article has no answer beyond the lag period, while the headline sells reassurance as a conclusion.

The Verdict

This is a lag snapshot masquerading as a systemic forecast. It documents AI entering Swiss SMEs without immediate upheaval; it does not show that human labor remains necessary. Under DT mechanics, the decisive question is not whether firms currently fear layoffs. It is whether AI progressively performs economically necessary cognitive work more cheaply and reliably than people. The article answers that question by measuring sentiment instead of structure. Its calm is not stability. It is the quiet phase of obsolescence.