AI boom poses new financial stability risks, BIS head says - Yahoo! Finance Canada

URL SCAN: AI boom poses new financial stability risks, BIS head says - Yahoo! Finance Canada
FIRST LINE: LONDON, Sept 10 (Reuters) - AI's rapid rise is creating new financial stability risks, Bank for International Settlements head Pablo Hernandez de Cos said, with spending on related infrastructure already at a scale significant enough to influence global economic conditions.

The Dissection

The article frames AI primarily as a financial-stability and productivity-management problem: opaque debt, private credit, trade concentration, infrastructure spending, productivity gains, and eventual reskilling. It accurately identifies the financial plumbing of the boom, but it treats labor displacement as a secondary adjustment cost rather than the central systemic rupture.

The article is watching the leverage around the machine while barely examining what the machine does to the human income circuit.

The Core Fallacy

It conflates economy-wide productivity growth with broad economic viability. A 10%–65% productivity gain on selected tasks, or a projected half-point increase in total-factor-productivity growth, does not preserve mass employment, wages, or purchasing power. It can mean fewer humans are required to produce more output.

Under the Discontinuity Thesis, the decisive sequence is not “AI raises productivity, workers retrain, growth broadens.” It is cognitive automation dominance, failed coordination to preserve human-only work, and collapse of productive participation. The article treats “retraining and reskilling” as a bridge without demonstrating that enough economically necessary human roles remain on the other side.

Hidden Assumptions

• Labor can be reallocated as quickly as AI can eliminate cognitive tasks.
• New jobs will appear at sufficient scale and quality to replace destroyed wage income.
• Productivity gains will be distributed rather than captured by AI-capital owners.
• Corporate earnings, debt, and private credit can absorb the infrastructure boom without a destabilizing correction.
• Opaque and interconnected financing is merely a risk to monitor, not evidence of a leveraged transition regime.
• Advanced economies can benefit first without intensifying ownership concentration and labor exclusion.
• Emerging economies can close the gap through digital infrastructure even if the underlying advantage shifts toward control of compute, energy, logistics, and models.
• “How widely benefits are shared” is treated as a policy variable rather than a conflict over ownership and power.

Social Function

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

The partial truth is real: AI investment is debt-linked, opaque, internationally concentrated, and capable of transmitting shocks through markets and trade. The transition-management function is to convert structural displacement into a familiar policy checklist—skills, infrastructure, adoption, and redistribution.

The anesthetic lies in presenting job losses as limited because the early phase is measured by visible layoffs rather than declining bargaining power, hiring avoidance, wage compression, and the disappearance of entry-level pathways. The system can be dying before the headline unemployment rate confesses it.

The article also performs elite self-exoneration. It acknowledges that benefits may not be widely shared while avoiding the ownership question: who controls the models, compute, data, energy, distribution channels, and resulting cash flows?

The Verdict

This is a credible financial warning wrapped around an incomplete social diagnosis. It correctly sees that the AI boom may be a leveraged, opaque asset-and-infrastructure cycle. It fails to recognize that the deeper threat is not merely an AI credit bust but the severing of the mass employment–wage–consumption circuit.

The BIS is measuring whether the machine destabilizes finance. The more terminal question is whether finance can remain socially legitimate once the machine makes most human cognitive labor economically unnecessary. Reskilling is not a solution if the system is manufacturing fewer economically necessary workers. The article identifies the smoke, catalogs the insurance exposure, and leaves the structural fire mostly unnamed.