Artificial intelligence and the challenge of oversight
AI is suddenly everywhere and it will change what you do, how you do it and what information will ultimately float to you, setting the stage for the decisions you make. Artificial intelligence has been compared to earlier technological transformations, from the steam engine to the internet. Yet the analogy oversimplifies. Previous transformations essentially reshaped how economies produce. The AI transition is also reshaping governance and it is increasingly about who gets to set the rules of the game.
AI appears to follow a familiar pattern with a revolutionary general-purpose technology spreading across sectors, boosting productivity and transforming labor markets. The concentration of economic power it produces, however, is not comparable to anything seen in earlier technological eras. A very small number of large technology firms control all the infrastructure required to develop and deploy advanced AI, hyperscale computing, vast proprietary datasets, specialized semiconductors, and frontier models.
That concentration is driven by increasingly steep barriers to entry and the network effects that characterize the digital economy. Training and operating advanced AI systems demands enormous fixed investments, and access to global-scale computing resources and data. As a result, AI is emerging not as a widely diffused technology but as one whose core capabilities are clustered in a few corporate and geographic hubs.
Much of the public debate has focused on the economic consequences of this concentration, the potential for job displacement, rising inequality, and widening gaps between technological leaders and laggards. These concerns are real. But the most profound effects of AI may ultimately prove political rather than purely economic.
When control over foundational technologies becomes concentrated, market power can evolve into rule-setting power. In earlier industrial eras, dominant firms wielded significant economic influence, but the authority to define market rules, competition policy, labor standards, financial regulation, remained largely in public hands. Today that boundary is becoming increasingly blurred.
Major technology platforms do not simply operate within regulatory systems; they actively shape them. As AI systems become embedded across finance, health care, education, public administration, and communication, the firms controlling foundational models and computing infrastructure acquire quasi-governance roles. Decisions about model design, access, pricing, and deployment carry broad social and economic consequences. In earlier eras, technological power translated into economic dominance. In the age of AI, it risks translating into governance capacity as well.
This dynamic produces a striking paradox. Debates with major implications for national security, economic competition, and information governance are increasingly shaped not by public institutions but by the strategic choices and rivalries of a small number of private technology firms.
This transformation is unfolding at a moment when many countries are already experiencing heightened polarization and growing distrust in political institutions. Over the past decade, economists and political scientists have documented how economic dislocation and the perceived loss of control contributed to the rise of populist movements. Trade shocks, automation, and the steady delegation of policymaking authority to unelected experts and bureaucratic agencies fostered a pervasive sense that key decisions affecting citizens’ lives were being made beyond democratic reach.
Artificial intelligence risks intensifying that perception. Like earlier waves of technological change, it generates significant distributional tensions, between high- and low-skill workers, leading and lagging regions, capital and labor. But it also introduces a more subtle source of political friction, the growing sense that economic and social outcomes are increasingly governed by technological systems controlled by corporate actors. The political economy of AI is not only about who gets richer, it is about who gets to decide.
Yet the relationship between technology firms and politics is more complex than simple opposition. In some respects, the structural incentives they face are beginning to align. Both dominant digital platforms and populist movements share a deep skepticism toward institutional constraints that slow decision-making and impose oversight.
Regulatory agencies, independent courts, and multilateral governance structures are routinely portrayed by populist leaders as obstacles to decisive political action. For large technology firms, those same institutional frameworks can appear as barriers to rapid scaling, market expansion, or data access and use. The motivations differ, but the logic converges, digital monopolies and populist politics alike tend to treat institutional constraints less as safeguards of democratic governance than as obstacles to speed, scale, and control. The trust in government is already low in many countries and in order not to fall behind, bureaucracy is seen as something negative.
In the U.S., several proposals associated with populist political agendas, from weakening independent regulatory agencies to reducing antitrust enforcement or limiting the scope of federal oversight over digital platforms, would effectively expand the operating space of large technology firms. Similar patterns can be observed elsewhere, where attacks on courts, regulators, or supranational institutions often have the side effect of eroding the institutional constraints that historically mediated the relationship between markets and democratic governance.
Digital platforms play a central role in shaping how information is produced, distributed, and monetized. Political actors increasingly rely on these infrastructures to mobilize support and communicate with voters, often bypassing traditional intermediaries such as parties, unions, and legacy media. AI is poised to deepen this transformation. Generative AI systems are beginning to reshape how text, images, and video are produced and circulated, blurring the boundary between authentic and synthetic information, and increasingly between the true and the false. Control over foundational models and computing infrastructure is rapidly becoming strategically significant not only for economic competition but for shaping political debate itself.
For Europe in particular, the challenge is acute. The continent has become a global leader in regulating digital markets but remains heavily dependent on external providers for the core infrastructures of AI. The governance of AI risks drifting toward a system in which private technological power increasingly defines the boundaries of democratic authority. Who will set the rules, the AI tech firms or the governments. In order to speed up tech development, it seems that they must take the lead, and the government should fill the function of enabler of this development. In more authoritarian regimes, this would include heavy government involvement, in the U.S. less regulation to expand business and take the lead, in Europe there are endless challenges due to heavy government regulations and a tradition of holding technology development back. It is a complex dilemma, but in the end of the day, the market leaders will develop where the tech firms have a relative free hand to develop and implement their systems.