Goldman Partner: AI Risks Creating 'Huge Danger' for Banking Skills | The Tech Buzz

TEXT START: Goldman Sachs is charging ahead with AI adoption across its operations, but a senior technology partner at the firm is sounding an alarm about an unintended consequence that could reshape the future of finance.

The Dissection

The text is laundering a structural labor shock into a training problem. It correctly identifies that AI removes the repetitive analytical work through which junior bankers traditionally acquired judgment. But it treats that loss mainly as a threat to human competence, not as evidence that the economic system no longer needs the same volume of human competence.

The article’s proposed remedy—manual modeling, AI-free assignments, and delayed access—preserves a ritual of professional formation after the market value of that ritual has begun to collapse. Goldman’s concern is operationally real: degraded human oversight can produce errors, liability, and reputational damage. It is not evidence that mass banking employment remains economically necessary.

The Core Fallacy

The central error is confusing “humans must retain enough judgment to supervise AI” with “large numbers of humans must perform the underlying work.” Those are not equivalent.

Under the Discontinuity Thesis, AI’s advantage is not merely faster information retrieval. It is the progressive automation of modeling, analysis, drafting, and reasoning—the exact tasks the article says build banker competence. Once AI becomes cheaper and more capable, firms have a competitive incentive to reduce the human layer rather than preserve it for educational purposes.

The article assumes AI will remain a tool that amplifies human judgment. The harsher possibility is that judgment itself becomes concentrated in a small ownership and control class, while most junior workers lose the productive pathway through which they might have become indispensable. The training ladder is not being repaired. Its lower rungs are being eaten.

Hidden Assumptions

• Banks will continue hiring large junior cohorts even after AI eliminates much of their billable workload.
• Competitive pressure will permit firms to sacrifice efficiency for human development.
• Human judgment will remain broadly superior enough to justify mass employment rather than narrow oversight roles.
• Manual practice can reproduce the strategic value of experience after AI has surpassed humans at much of the underlying analysis.
• The problem is primarily skill atrophy, rather than the collapse of productive participation for ordinary knowledge workers.
• Institutional guardrails can preserve human-only economic domains at scale.
• The future banker will be a broadly distributed professional class, rather than a smaller population of owners, controllers, accountable signatories, relationship managers, and technical servitors.

These assumptions are not demonstrated. They are required to keep the article’s preferred conclusion intact.

Social Function

This is partial truth combined with transition management and elite self-exoneration.

The text admits that AI can hollow out the talent pipeline, which is a genuine warning. But it frames the crisis as something firms can solve through better onboarding and “guardrails,” allowing Goldman and its peers to present themselves as responsible stewards of human expertise while continuing the automation race.

Its ideological function is to preserve the comforting distinction between AI as a tool and AI as a replacement. The article never seriously confronts the competitive endpoint: if a machine can perform the junior work, retaining humans to perform it for training purposes becomes an expensive luxury. The institution may preserve a thin layer of human judgment because it needs accountability and exception handling. That does not preserve the old labor market.

The Verdict

The article diagnoses the symptom and misnames the disease. Banker judgment is not merely at risk of erosion; the economic pipeline that produced and rewarded mass banker judgment is becoming obsolete.

The proposed manual-training safeguards are hospice care for a disappearing apprenticeship system. Goldman may still need elite Sovereigns and indispensable Servitors—owners, controllers, relationship brokers, regulators, signatories, and specialists who manage AI, capital, energy, logistics, and institutional liability. It will not need the previous army of junior analysts in anything like its former numbers.

The warning is therefore not a case for “human expertise while leveraging machine efficiency.” It is an early admission that AI is severing the labor-to-status-to-consumption circuit from inside one of its most prestigious institutions.