"A warning sign we shouldn't ignore": Thousands of jobs gone as AI hits Irish - IrishCentral

TEXT START: The finding was noted in a Labor Force survey earlier this year, which showed a drop of 20,000 jobs in the information and communications technology (ICT) sector, and was shared with Mr Harris as Finance Minister.

The Dissection

This article documents an early fracture, then surrounds it with institutional sedation. The concrete signal is severe: ICT employment fell by 20,000 year-on-year, including 16,200 jobs in computer programming and consultancy; AI-exposed sectors are growing more weakly; younger workers are losing entry-level routes. The rest is damage control: survey volatility, strong aggregate employment, 250 new OpenAI jobs, reskilling, redeployment, and a five-to-ten-year implementation horizon.

The article is not really asking whether AI is replacing labor. Its function is to make replacement appear administratively manageable. It converts a structural employment problem into a skills-policy problem.

The Core Fallacy

The central error is confusing aggregate employment and productivity gains with continued mass productive participation.

A record 2.84 million people at work does not disprove displacement. It can coexist with AI eliminating the first rung of the career ladder while labor is temporarily absorbed elsewhere. Likewise, 250 new jobs do not compensate for 20,000 lost ICT positions. They are a showcase number, not a labor-market replacement mechanism.

The article also assumes that workers can simply “upskill” into the new economy. That ignores the competitive mechanism: when AI raises the output of skilled workers, firms need fewer of them. Better skills may improve an individual’s ranking while shrinking the number of economically necessary positions. The scarce resource is not human capability. It is human necessity.

Under the Discontinuity Thesis, this is the opening phase of P1 and P3: cognitive work becomes cheaper to automate, and younger workers lose access to economically necessary labor before aggregate employment statistics visibly collapse.

Hidden Assumptions

• New AI-created jobs will appear at sufficient scale to replace destroyed jobs.
• Displaced workers can transition into those jobs without a prolonged exclusion period.
• The skills that make Ireland attractive to AI firms will remain scarce after AI adoption spreads.
• Strong tax receipts and total employment indicate durable labor demand rather than lagging macroeconomic inertia.
• Public-sector redeployment restores productive participation rather than merely hiding redundant labor inside the state.
• A five-to-ten-year adjustment period is a solution rather than a delay before the institutional bill arrives.
• Entry-level workers can gain experience without the entry-level work that formerly supplied it.
• AI adoption will enhance productivity broadly without concentrating ownership and bargaining power.

These assumptions all treat the economy as a machine that automatically manufactures replacement roles. The article provides no mechanism proving that it will.

Social Function

Primary classification: transition management and ideological anesthetic, with a substantial partial-truth component.

The article admits the warning sign because denial would be visibly absurd. It then neutralizes the warning through familiar institutional assurances: monitor trends, reskill workers, support firms, capture opportunities, redeploy personnel, and wait several years. This preserves confidence in the existing policy machinery while the underlying labor circuit weakens.

The rhetoric also performs elite self-exoneration. If workers fail, the implied cause will be inadequate adaptation. The structure escapes blame; the displaced individual is assigned a skills deficit. “Ireland is well positioned” becomes a way to say that the country may be first in line for automation while pretending that proximity to the frontier is protection.

The Verdict

This is a partial report of systemic failure wrapped in a manageable-transition narrative. The 20,000-job fall—especially the 16,200 reduction in programming and consultancy and the damage to younger workers—is the meaningful evidence. The employment record, tax receipts, and 250-job announcement are lagging indicators and public-relations counterweights, not refutations.

The article correctly identifies the first casualties but mislabels the mechanism as a temporary skills mismatch. The deeper process is labor devaluation: AI attacks the cognitive entry routes that reproduce the workforce, then makes each surviving worker more productive and therefore less numerous. Ireland may capture AI investment. That does not mean Irish workers will capture the income or necessity created by it.

The post-war employment bargain is not yet visibly dead in the statistics. It is already being hollowed out at the point where new workers are supposed to enter it.