Will the AI boom increase inequality?

Sunday 29 March 2026

The modern enthusiasm for artificial intelligence is often framed in the language of progress – efficiency, productivity, discovery. Yet beneath this rhetoric lies a more disquieting question, one that echoes across centuries of industrial transformation: who, precisely, will benefit? In his most recent annual letter Larry Fink, the chief executive of BlackRock, has given this question renewed urgency, warning that the boom in artificial intelligence risks deepening the already profound inequalities that define contemporary capitalism.

Fink’s concern is neither novel nor alarmist. Rather it is grounded in a sober reading of economic history. The great technological revolutions – from the mechanisation of textile production to the rise of the internet – have seldom distributed their gains evenly. Instead they have tended to reward those already positioned to exploit them: those with capital, infrastructure, or privileged access to markets. In Fink’s words, “the massive wealth created over the past several generations flowed mostly to people who already owned financial assets”, and artificial intelligence threatens to replicate this pattern “at an even larger scale”. 

The structure of the contemporary AI economy reinforces this tendency. Unlike earlier industrial innovations, which often diffused gradually across firms and regions, artificial intelligence is capital-intensive in an extreme sense. It requires vast quantities of data, specialised computing hardware and highly trained personnel. These are resources concentrated in a small number of firms – predominantly large American and Chinese technology companies – and, crucially, in the investors who finance them. As a result the early gains of the AI revolution have accrued disproportionately to a narrow segment of society. The extraordinary rise in the valuation of companies such as Nvidia, now measured in trillions of dollars, is emblematic of this concentration of wealth. 

Fink’s argument extends beyond corporate structure to the distribution of financial ownership itself. Even in advanced economies, where participation in capital markets is widespread in principle, it remains highly unequal in practice. In the United States for example the majority of equities are held by a small fraction of households. If artificial intelligence drives a new wave of stock market gains – as it already appears to be doing – then those gains will flow primarily to those who already possess substantial financial assets. The result is what economists have described as a “K-shaped” economy, in which the trajectories of wealth and income diverge sharply between the affluent and the rest.

There is, however, a further dimension to the inequality question, one that concerns labour rather than capital. Artificial intelligence is not merely a tool of productivity; it is also a substitute for human work. While earlier technological changes tended to displace manual labour, AI increasingly encroaches upon cognitive and professional occupations. Fink has warned that even highly educated young people may face deteriorating employment prospects as entry-level roles are automated or transformed. This introduces a paradox: at the very moment when education has become the principal means of social mobility, the returns to that education may be diminishing.

Yet it would be simplistic to conclude that artificial intelligence must inevitably exacerbate inequality. History also provides counterexamples in which the benefits of technological change were broadly shared. The post-war decades in Western Europe and North America, for instance, witnessed rapid technological progress alongside declining inequality. The crucial difference lay not in the technologies themselves, but in the institutional frameworks that governed their deployment – progressive taxation, strong labour organisations, public investment in education, and policies designed to widen access to capital.

Fink’s own proposed remedy reflects this historical lesson, though from the perspective of a financial capitalist rather than a social democrat. He advocates for broader participation in capital markets, arguing that individuals must be enabled to share in the wealth generated by AI through long-term investment. This is a striking position, insofar as it implicitly acknowledges that wages alone will not suffice to distribute the gains of technological progress. Instead ownership – of shares, of funds, of financial instruments – becomes the central mechanism through which individuals may secure their economic future.

However this solution raises its own difficulties. Access to capital markets presupposes the ability to save, and for many households this remains an elusive condition. As Fink himself has observed, a significant proportion of individuals lack even modest financial reserves, rendering sustained investment impracticable. Moreover expanding participation in financial markets does not necessarily alter the underlying distribution of power within those markets. The concentration of ownership may be mitigated, but not fundamentally transformed.

The geopolitical dimension of artificial intelligence further complicates the picture. AI is not merely an economic phenomenon; it is also an instrument of strategic competition between states. Those countries able to dominate AI development will command not only economic advantages but also military and informational power. This introduces the prospect of inequality not only within societies but between them. Smaller or less technologically advanced states may find themselves increasingly dependent upon the AI capabilities of others, with profound implications for sovereignty and development.

In this respect the question of inequality becomes inseparable from that of governance. Who controls the data upon which AI systems are trained? Who owns the infrastructure that sustains them? Who determines the rules under which they operate? These are not merely technical questions; they are political ones, and their answers will shape the distribution of wealth and power in the decades to come.

The boom in artificial intelligence then does not carry an inherent destiny of inequality. Rather it amplifies existing structures, rewarding those already positioned to benefit while exposing the vulnerabilities of those who are not. Fink’s warning should therefore be understood not as a prophecy but as a diagnosis. Artificial intelligence will increase inequality if the institutions surrounding it remain unchanged. It will not do so if those institutions are reformed to ensure broader participation in its gains.

The challenge, as ever, lies in the will to act.