Rawls himself acknowledged the force of the incentive argument. The difference principle permits inequalities that serve as genuine incentives for productive activity, provided the inequalities actually benefit the least advantaged. If outsized returns to AI investment are necessary to motivate the investment, and if that investment produces gains that ultimately improve the condition of the worst-off, then the outsized returns are permitted. The principle does not demand equality of outcome; it demands that inequality serve a purpose, and that the purpose be the benefit of those at the bottom.
But the dynamic efficiency objection proves less than its proponents claim. It assumes that the current level of inequality is the minimum necessary to incentivize productive activity — an empirical claim that the objection rarely establishes. It assumes that technology companies and their investors would cease to innovate under moderately progressive taxation — another empirical claim for which the historical evidence is weak. It assumes that the market, left to its own dynamics, distributes gains in a way that no deliberate institutional design could improve upon — a claim that the history of technological transitions contradicts.
The historical record, as documented by economists Daron Acemoglu and Simon Johnson in Power and Progress, demonstrates that the gains of technology are not automatically shared. They require institutional intervention — labor movements, legislation, regulatory frameworks — to translate into broadly distributed improvements in living standards. The labor protections of the early twentieth century did not end industrialization. The environmental regulations of the late twentieth century did not end chemical manufacturing. In each case, institutional intervention produced a period of adjustment — during which the affected industries protested, predicted catastrophe, and eventually adapted — followed by a period of innovation within the new constraints that proved as productive as the innovation that preceded it.
G.A. Cohen's critique of the incentive argument sharpens the Rawlsian response. Cohen argued that if the talented could produce the same output without inequality-generating incentives — if the surgeon would perform surgery at a lower wage, if the technology executive would innovate at a lower return — then the inequalities are not justified by the difference principle; they are justified only by the fact that the talented hold their labor hostage to extract rents that the basic structure permits. Cohen's critique insists that the principle be applied rigorously: Is this inequality the minimum necessary to secure the gains that benefit the least advantaged? Or is it rent extraction disguised as incentive? Most current AI industry inequalities, evaluated against this standard, are more plausibly the latter than the former.
The dynamic efficiency objection in its modern form descends from Nozick's critique of Rawls in Anarchy, State, and Utopia (1974) and from the broader neoclassical economic critique of redistributive policy. In the AI context, the objection has been advanced in various forms by technology industry spokespersons, libertarian commentators, and economists concerned about the effects of regulation on innovation incentives.
Incentive structures matter. The objection contains the genuine insight that institutional design must account for how rules shape productive activity.
Empirical claims underlying the objection. The objection depends on empirical claims about minimum necessary incentives that are rarely established and often implausible.
Historical precedent against. Previous technological transitions demonstrate that institutional intervention does not suppress innovation; it redirects it.
Cohen's critique of rent extraction. The difference principle, applied rigorously, asks whether inequalities are genuinely necessary for productive gains or merely permitted rent extraction.
Asymmetric burden of proof. The objection treats the current distribution as the baseline against which interventions are evaluated, but the current distribution itself requires justification under the framework.