For more than a decade, the financial services industry has braced for an existential reckoning that never quite arrived. Fintech would render traditional banks obsolete. JPMorgan, HSBC, and their peers would be hollowed out by nimble digital challengers. BigTech would absorb banking altogether. Digital wallets would cut banks out of the payment chain entirely. Cryptocurrencies would replace the deposit system. Stablecoins would render central-bank money redundant. Decentralised finance — DeFi — would eliminate intermediaries altogether. Wave after wave of predicted disruption crashed against the institutional architecture of global banking, and the architecture held. Now, according to Chris Skinner writing on The Finanser, artificial intelligence represents a fundamentally different kind of challenge — one that may finally deliver what its predecessors could not.
The pattern Skinner identifies is worth dwelling on. Each successive disruption narrative followed a remarkably similar arc: a genuinely transformative technology emerges, commentators declare the end of traditional banking, incumbent institutions scramble, regulators convene task forces, and then — over the course of years — the feared displacement either fails to materialise or takes a far more constrained form than prophesied. Fintech produced valuable niche players, many of which eventually partnered with or were acquired by the banks they were meant to destroy. Cryptocurrency markets ballooned and collapsed in alternating cycles without displacing fiat deposit systems. Decentralised finance attracted billions in speculative capital but remained structurally peripheral to the mainstream financial system.
This track record of non-disruption has bred a certain institutional complacency — and perhaps reasonably so. Bank chief executives who have watched five or six supposedly epoch-defining technology waves break harmlessly against their balance sheets could be forgiven for treating each new alarm with measured scepticism. The danger, of course, is that this habituation to false alarms leaves institutions systematically under-prepared for the wave that finally carries real consequence.
Skinner's framing suggests that artificial intelligence may be precisely that wave. The distinction worth making is structural rather than rhetorical. Previous disruption candidates — DeFi, stablecoins, digital wallets — were largely additive technologies. They created new channels, new instruments, or new asset classes, but they did not fundamentally alter the cognitive and operational infrastructure through which financial decisions are made. Artificial intelligence is different in kind. It does not merely create a new product category; it rewrites the logic by which risk is assessed, credit is allocated, fraud is detected, trading decisions are executed, and customer relationships are managed. When AI begins to perform the analytical and decisional functions that have historically required human expertise within regulated institutions, the question of who is "running" the financial system becomes genuinely complicated.
The implications for regulation are profound and underexplored. The European Banking Authority (EBA), the European Central Bank (ECB), and their counterparts across major jurisdictions have begun grappling with AI governance frameworks, but the speed of deployment in financial services is outpacing the pace of supervisory adaptation. When a large language model or a proprietary AI system is effectively making credit decisions or flagging transactions, accountability becomes diffuse in ways that existing regulatory architecture was not designed to handle. The chartered bank may still hold the licence, but if the cognition is outsourced to a technology vendor's model, the practical locus of financial decision-making has migrated outside the regulatory perimeter.
This is precisely the concern that Skinner's question — are tech firms now running our financial systems? — is designed to surface. It is not a question about ownership or formal control. Banks have not been acquired by Silicon Valley. Licences remain with regulated institutions. But operational dependence on technology infrastructure controlled by a small number of large technology companies creates a form of functional dominance that does not require formal ownership to be consequential. If a single AI platform processes the risk assessments for a significant portion of a national banking system, the failure or compromise of that platform becomes a systemic risk — one that regulators who focus on individual institution solvency may be poorly positioned to detect or contain.
What This Means for the Industry
The honest assessment is that the banking industry has been correct, so far, to resist the more extreme disruption forecasts of the past decade. Fintech did not destroy banking. DeFi did not replace everything. Stablecoins have not displaced deposits. That record of institutional resilience is real and should not be minimised. But resilience against yesterday's disruption vectors is not evidence of resilience against tomorrow's. Artificial intelligence operates at a different layer of the financial stack — not the product layer or the channel layer, but the decisional layer — and it is precisely at that layer that the boundary between a bank running its own operations and a technology firm running those operations on the bank's behalf becomes genuinely unclear. Regulators, boards, and investors would be well served to treat that ambiguity not as a philosophical curiosity, but as an emerging material risk that demands structural answers before the next supervisory cycle begins.
Written by the editorial team — independent journalism powered by Codego Press.