Bridgewater: AI Could Displace 18% of US Jobs

By Newsmax Wires    |   Monday, 17 August 2026 08:42 AM EDT

Artificial intelligence could displace 18% of current U.S. jobs within the next five years, according to an internal analysis by Bridgewater Associates cited in a New York Times opinion piece published Friday.

The warning comes from Greg Jensen and Nir Bar Dea, two senior executives at the investment firm, who argue that the United States faces a critical policy choice: manage AI-driven disruption now or risk a severe backlash later.

In their Aug. 14 column, “This Is One of the Most Important Policy Decisions of Our Lifetime,” Jensen and Bar Dea present AI as both an extraordinary economic opportunity and a potentially destabilizing force.

They contend that AI could produce a “productivity miracle” capable of substantially improving living standards, but only if governments establish policies that distribute its gains broadly and limit its dangers.

Jensen is Bridgewater’s managing chief investment officer, while Bar Dea serves as the firm’s chief executive officer.

Bridgewater, the $102 billion investment firm they lead, has invested heavily in artificial intelligence and reorganized parts of its business to capitalize on the technology.

That position gives their argument an unusual dimension: The policies they recommend could impose costs on their own company and the investment interests they represent.

The authors’ central concern is not simply that individual workers will lose their jobs.

They fear that widespread displacement could undermine social stability if the economic benefits of AI become concentrated among companies, investors, and highly skilled workers.

They point to historical episodes following major economic transformations, including the political upheavals associated with industrialization, as evidence that periods of rapid growth can produce dangerous social tensions when wealth and opportunity become too concentrated.

Jensen and Bar Dea say their internal analysis estimates that 18% of existing U.S. jobs could be displaced by AI over the next five years.

They acknowledge that new jobs will emerge, particularly occupations dependent on human relationships, such as nursing and hospitality.

Nevertheless, they argue that the transition itself could be socially disruptive.

Their most distinctive proposal is an AI token tax.

AI systems process information in units known as tokens, and companies using enterprise AI models often pay according to the number of tokens consumed.

The authors describe token consumption as roughly analogous to wages paid for machine labor.

Because traditional taxes fall primarily on human labor, they argue, the current tax system gives businesses an incentive to substitute machines for people.

Revenue from a token tax, they propose, could be used to reduce taxes on human labor while simultaneously financing a national ownership program for AI companies.

Under their proposal, the government would acquire stakes in major American AI firms and distribute those shares directly to citizens.

Rather than allowing the government to retain ownership, individuals would become shareholders and benefit financially from AI-driven economic growth.

The authors also call for significantly tougher AI safety rules.

They argue that regulation should cover not only the public release of advanced models but also their development and use.

Their concern is that increasingly autonomous systems could develop harmful capabilities, act beyond their intended constraints, or be repurposed in dangerous ways.

Jensen and Bar Dea conclude that policymakers have only a limited window to act.

Their message is deliberately urgent: America should encourage AI innovation while ensuring that workers and citizens share in its rewards.

Without that balance, they warn, an economic revolution that could raise prosperity for millions could instead deepen inequality, provoke political backlash, and create risks that become much harder to control once AI is widely deployed.

Categories: Articole de interes general