- Can banks withstand AI concentration?
- The bigger risk isn’t finance
- Why workforce disruption matters most
The debate surrounding artificial intelligence (AI) often gravitates toward dramatic scenarios: market disruptions, algorithmic failures, and the possibility of financial instability triggered by increasingly autonomous systems. Such concerns are understandable. For regulators, central banks, and financial institutions, memories of the 2008 global financial crisis remain deeply embedded in institutional thinking. Any emerging technology capable of amplifying systemic vulnerabilities naturally invites heightened scrutiny.
Yet, while financial-system risks associated with AI deserve careful monitoring, an increasingly persuasive argument is emerging among economists and policymakers: the more consequential challenge may not lie within banking systems themselves, but within labor markets and broader economic structures. Put differently, the pressing question is not whether AI will destabilize finance, but whether economies are adequately prepared for its impact on employment, skills, and income distribution.
For banking leaders across the Arab world, this distinction matters. It influences regulatory priorities, workforce planning, investment strategies, and the role financial institutions themselves may play in supporting economic resilience.
Lessons from 2008: A Stronger Financial Safety Net
To understand why an AI-induced financial shock may be more manageable than often assumed, one must first revisit the legacy of the 2008 financial crisis.
The collapse of major financial institutions—most notably the failure of Lehman Brothers—exposed profound weaknesses in the architecture of global finance. Hidden leverage, opaque counterparty exposures, fragmented supervision, and insufficient systemic oversight created conditions in which the failure of one institution threatened to cascade through the broader financial system.
In the aftermath, regulators across the United States, Europe, and major global markets embarked on one of the most extensive financial reform agendas in modern history. Capital requirements were strengthened. Resolution frameworks were enhanced. Supervisory coordination improved. More importantly, regulators shifted their mindset from overseeing individual institutions in isolation to assessing interconnected systemic risk.
One of the most consequential outcomes of these reforms was the recognition that certain institutions had become “too interconnected to fail.” Rather than waiting for collapse and improvising emergency rescues, regulators introduced mechanisms designed to ensure that systemically important institutions could withstand severe shocks—or fail in a controlled manner without destabilizing the broader system.
This regulatory maturity has important implications for AI.
While the technology itself differs significantly from financial instruments, the governance logic developed after 2008 offers a useful framework for assessing emerging vulnerabilities associated with concentrated AI infrastructure and dependencies.
A Concentrated AI Ecosystem
The parallels between today’s AI sector and pre-2008 finance are difficult to ignore.
Much like the banking system before the global crisis, the AI ecosystem exhibits significant concentration. A limited number of companies dominate advanced large-language-model capabilities, cloud infrastructure, and semiconductor production.
In enterprise-grade generative AI, a small cluster of technology firms currently controls the overwhelming majority of market activity. Equally significant is the concentration at the hardware level, where advanced semiconductor manufacturing is heavily dependent on a narrow set of suppliers.
This concentration introduces a new form of systemic exposure.
If major AI providers become deeply embedded in financial institutions’ risk management, fraud monitoring, customer services, compliance systems, treasury operations, or investment analysis, operational dependence on a small number of technology providers may become increasingly difficult to ignore.
The concern is not merely commercial concentration. It is infrastructural concentration.
A single disruption—whether technological, geopolitical, regulatory, or environmental—could have cascading operational implications across multiple jurisdictions and financial systems simultaneously.
The temporary disruption to semiconductor production in Taiwan following the April 2024 earthquake served as a reminder of how geographically concentrated parts of the global technology supply chain have become. In an era where financial institutions increasingly rely on computing power and AI-enabled systems, resilience of underlying infrastructure deserves greater strategic attention.
For Arab central banks and financial regulators, the issue is particularly relevant given the region’s accelerated investments in digital banking, fintech ecosystems, and cloud-based financial infrastructure.
Why the Financial Risks May Be Contained
Despite these concerns, there are strong reasons to believe that AI-related financial risks can remain manageable—provided regulators act proactively and institutions strengthen operational preparedness.
Unlike 2008, regulators are not approaching this challenge without precedent or institutional memory.
Global supervisory bodies already possess tools for identifying critical dependencies, assessing concentration risk, and developing resilience requirements. What may be required is adaptation rather than reinvention.
Several priorities deserve particular attention.
First: Identifying Systemically Important AI Infrastructure
Financial supervisors may increasingly need to identify which technology providers have become essential to financial stability.
This extends beyond AI developers themselves to include cloud-service providers, chip manufacturers, and digital infrastructure operators whose failure could disrupt critical financial services.
The challenge today lies partly in visibility. Regulatory frameworks for monitoring AI-related dependencies remain at an early stage, and meaningful data gaps persist. Without standardized reporting or common classifications, authorities may struggle to understand where concentrated exposures truly exist.
For financial institutions in the Arab region, this suggests a growing need for internal mapping of technology dependencies, particularly where critical functions rely heavily on external AI systems.
Boards and senior management should increasingly ask a straightforward but strategic question: What happens if our primary AI provider becomes unavailable?
Operational Resilience Must Become a Strategic Priority
The second imperative concerns resilience.
In traditional banking supervision, capital buffers help absorb losses during crises. AI systems, however, do not operate like balance sheets. Their vulnerabilities lie in operational continuity, model availability, data access, and infrastructure reliability.
As such, resilience may increasingly become the functional equivalent of financial capital.
Institutions relying heavily on AI should consider redundancy arrangements, diversified providers, backup systems, and contingency planning. Overreliance on a single provider—particularly for mission-critical functions—may create concentration risks similar to excessive counterparty exposure in finance.
This consideration becomes especially relevant as Arab banks continue expanding digital offerings, automated compliance monitoring, customer analytics, and AI-assisted operational processes.
The region’s banking sector has made considerable progress in digital adoption. Yet sophistication must be accompanied by preparedness. Technological efficiency should never come at the expense of institutional resilience.
Stress Testing in the Age of AI
Another lesson from post-crisis banking is the importance of stress testing.
For years, financial stress tests have served as an essential mechanism for identifying vulnerabilities before crises materialize. Similar approaches may increasingly become relevant in assessing AI-related dependencies.
What would happen if a dominant cloud provider experienced prolonged disruption? How resilient are institutions to sudden regulatory restrictions affecting major AI systems? Could geopolitical tensions interrupt semiconductor supply chains critical to digital infrastructure?
These are no longer theoretical questions.
AI systems are fundamentally shaped by historical data, which creates limitations in responding to highly unusual or unprecedented events. Financial institutions therefore cannot assume that AI-enabled systems will necessarily perform effectively under conditions of severe uncertainty.
Judgment, governance, and human oversight remain indispensable.
The Larger Economic Question: Employment and Workforce Disruption
Yet even if financial-system vulnerabilities are effectively contained, a more profound issue remains: what AI may mean for workers, productivity, and social cohesion.
Here, the implications become considerably broader.
Evidence of workforce disruption is already emerging across sectors. Large corporations have begun reducing hiring in functions where automation increasingly handles routine or repetitive tasks. Human resources, customer service, administrative processing, coding support, and operational analysis are among the functions seeing accelerated automation.
Forecasts suggest that a meaningful share of existing workplace skills could require significant adjustment within the next decade.
For Arab economies—many of which already face youth-employment pressures, skills mismatches, and labor-market rigidities—this issue warrants serious attention.
The challenge is not necessarily mass unemployment. Rather, it is the pace of adaptation.
Historically, technological progress has created new jobs even as older roles declined. The concern today is whether reskilling systems can evolve quickly enough to support workers transitioning into new forms of employment.
This is particularly relevant for economies in the Middle East and North Africa, where demographic realities make workforce inclusion a strategic economic priority.
A Shared Responsibility Between Governments and Institutions
Addressing workforce disruption cannot be left solely to governments.
Employers, educational institutions, regulators, and financial institutions all have a role to play.
Training initiatives will matter—but effectiveness depends on alignment with real market demand rather than generic skills programs. Workers must be equipped with capabilities that are commercially relevant, technologically adaptable, and economically valuable.
Equally important is digital access. Broadband infrastructure, digital literacy, and affordable technological access increasingly represent economic infrastructure—not optional public policy considerations.
Banks themselves may become key enablers of this process, whether through financing education initiatives, supporting SME adaptation, investing in workforce development, or enabling financial inclusion linked to digital skills.
The broader economic objective should be clear: ensuring that the gains associated with AI contribute to wider prosperity rather than becoming concentrated among a narrow segment of technology owners.
A Strategic Imperative for Arab Banking Leaders
For Arab banking leaders, the conversation around AI should move beyond either excessive optimism or alarmism.
The immediate systemic financial risks appear manageable, particularly if regulators and institutions apply lessons learned from the post-2008 era. Concentration risk, operational resilience, and stress testing should become part of mainstream governance discussions surrounding AI adoption.
At the same time, leaders should avoid viewing AI solely through a productivity lens.
The more consequential challenge may lie in employment, workforce preparedness, and economic inclusion. Banks cannot remain passive observers in this discussion. As key intermediaries between capital, business activity, and national development priorities, financial institutions will inevitably influence how economies adapt.
The institutions best positioned for the coming decade will likely be those that pair technological sophistication with prudent governance, invest in workforce readiness, diversify operational dependencies, and maintain a balanced view of both opportunity and risk.
In the end, the question is not whether AI will influence finance and economies—it already is. The more important question is whether institutions are preparing thoughtfully enough to ensure that the benefits are broadly shared, operational risks remain manageable, and long-term economic stability is strengthened rather than tested.
