Who profits from banking's AI moment? - EUobserver

TEXT START: Generative artificial intelligence has arrived in Europe’s banks faster than almost anyone predicted, and it is already changing how the sector’s 2.6 million workers do their jobs.

The Dissection

This is a stakeholder defense brief disguised as balanced analysis. It admits the essential facts: AI adoption is widespread, it creates enormous value, workers fear displacement, and Intesa Sanpaolo is eliminating 9,000 jobs while hiring only 3,500 digitally skilled replacements.

The article then relocates the problem from automation to consultation. Its central move is to present social dialogue, retraining and human oversight as mechanisms capable of making labor displacement fair. The paid sponsorship matters: the piece normalizes rapid AI deployment while preserving the legitimacy of unions, employers and regulators as managers of the transition.

The Core Fallacy

It confuses management of obsolescence with prevention of obsolescence.

Under the Discontinuity Thesis, once AI performs fraud checks, compliance reporting, credit analysis and customer support at lower cost, competitive pressure forces banks to capture those savings. Unions can negotiate timing, severance, redeployment, monitoring and procedural safeguards. They cannot make human labor economically necessary again.

The article’s claim that AI can make jobs “less routine, more human” is a temporary description of the transition layer. As systems improve, the remaining human tasks become narrower, more supervised and more vulnerable to the next automation wave. Human oversight itself becomes a cost center to be minimized.

The result is not preservation of the postwar employment-to-consumption circuit. It is its administrative liquidation, padded with dialogue.

Hidden Assumptions

• Productivity gains will become lower workloads or better services rather than permanent headcount reduction, despite the article’s own Intesa example.
• New digital hires are equivalent to eliminated workers, despite the obvious skills and status mismatch.
• Retraining can work at scale for older and less digitally literate employees—the groups the article identifies as most exposed.
• The reported 200,000 jobs at risk captures the full displacement rather than only visible redundancies.
• EU regulation can constrain substitution rather than merely document, monitor and legitimize it.
• Collective bargaining coverage will expand from roughly 20 percent before automation accelerates further.
• Workers can meaningfully govern AI systems without controlling the capital deploying them.
• Banks under competitive pressure will voluntarily leave efficiency gains as reduced workloads instead of harvesting them as labor savings.
• Consultation can repair distrust without changing the underlying loss of bargaining power.

Social Function

Primary classification: transition management.

Secondary classifications: partial truth, ideological anesthetic, elite self-exoneration and stakeholder propaganda. The article is not fabricating the disruption. It is laundering its meaning. A capital-driven substitution process is reframed as a jointly governable social project, allowing banks to claim responsibility and unions to claim influence while the economic necessity of human labor contracts.

“Fair transition” here means a more orderly distribution of casualties. It does not mean that the old system survives.

The Verdict

The article correctly identifies banking’s first AI incision and then draws the wrong systemic conclusion. The profits accrue to the banks and those controlling their capital; workers receive, at best, negotiated delay, retraining and improved redundancy terms.

Social dialogue can make the purge slower and less brutal. It cannot defeat P1, P2 or P3. Banking’s AI moment is a profit moment for capital and a managed-obsolescence moment for labor. The article’s humane transition is hospice care for the employment circuit, complete with consultation rights and a timetable.