Major technological shifts tend to produce major social upheaval, and artificial intelligence is no exception. The scale of what’s at stake, however, cuts both ways: AI could deliver a genuine productivity leap capable of transforming living standards, but only if governments move now to head off its worst consequences before they arrive.
That warning comes, notably, from within the industry benefiting most from AI’s rise. Bridgewater, the asset management firm, has spent years building its business around AI, restructuring itself to extract as much value from the technology as possible, a position that means it stands to be disproportionately affected by the very taxes and regulations its leadership is now recommending. That the recommendation comes from a firm with skin in the game, rather than against its own interest, is part of what makes the argument worth taking seriously.
The core claim is straightforward: effective AI policy has to accomplish two things simultaneously, building society’s resilience to disruption and averting outright technological catastrophe. Miss either target, and the promise of an AI-driven productivity boom falls apart. Whether such a boom ends up benefiting society broadly, the argument goes, depends less on the technology itself and more on whether governments can hold the social fabric together through the transition, preserving a sense of fairness and shared stake in the outcome. History offers a blunt warning about what happens when that fails: wealth concentration paired with rising disaffection has a track record of ending badly, from the rise of Communism and fascism following the Industrial Revolution to the populist backlash against globalization playing out today.
Much remains uncertain about exactly how AI-driven disruption will unfold, but the scale of the shift already looks significant. Internal analysis from Bridgewater suggests roughly 18 percent of current American jobs could be displaced by AI within five years. New jobs will emerge too, particularly in fields where human connection is central to the work itself, such as nursing or hospitality, but the broader disruption from this transition appears largely unavoidable regardless. The argument is that governments need to act ahead of the backlash this labour displacement will eventually provoke, not scramble to respond once it has already arrived.
One concrete recommendation centers on tax policy: current systems tax human labour while leaving machine labour largely untouched, tilting the incentive structure toward automation by default. A consumption tax applied to AI usage itself, the argument suggests, could help correct that imbalance.
