Artificial intelligence could push nearly one in five existing American jobs out of the workforce within the next five years, according to an internal analysis from investment giant Bridgewater Associates, whose top executives are warning that Washington must act quickly to prevent the technology’s economic promise from triggering major social upheaval.
The warning was issued by Bridgewater executives Greg Jensen and Nir Bar Dea, who say the United States is approaching a pivotal moment in determining how to handle the enormous changes expected from AI. They argue that policymakers can either prepare for the disruption now or face potentially severe public and political consequences later.
Jensen and Bar Dea laid out their case in an Aug. 14 New York Times opinion column titled “This Is One of the Most Important Policy Decisions of Our Lifetime,” describing artificial intelligence as a technology capable of generating enormous prosperity while simultaneously posing serious risks to economic and social stability.
The executives believe AI could unleash a “productivity miracle” that dramatically increases economic output and raises living standards. But they maintain that achieving that outcome will depend on government policies designed to spread the financial benefits of the technology across society while guarding against its potential hazards.
Jensen serves as Bridgewater’s managing chief investment officer, while Bar Dea is the investment firm’s chief executive officer.
The two executives oversee Bridgewater, a $102 billion investment firm that has itself made significant investments in artificial intelligence and reshaped portions of its operations in an effort to take advantage of the rapidly advancing technology.
That makes their recommendations particularly noteworthy, since some of the policies they are advocating could create new expenses for Bridgewater itself as well as for the investment interests represented by the firm.
Their warning extends beyond the immediate prospect of individual Americans finding themselves unemployed because their work can be performed by artificial intelligence.
Instead, Jensen and Bar Dea are concerned that mass job displacement could threaten broader social cohesion if most of the wealth created by AI flows primarily to corporations, investors and workers possessing highly specialized skills.
They cite earlier periods of dramatic economic change, including the upheaval that accompanied industrialization, to argue that extraordinary economic growth can also fuel intense political and social unrest when its benefits and opportunities are disproportionately concentrated.
According to Jensen and Bar Dea, Bridgewater’s internal research suggests that AI could displace 18% of the jobs currently held by American workers over the coming five years.
The executives concede that technological change will also create new employment opportunities, particularly in fields built around direct human interaction and relationships, including nursing and hospitality.
Even with those new jobs, however, they caution that the period of adjustment between the old economy and the emerging AI-driven one could itself produce significant social disruption.
Among their most unusual policy recommendations is the creation of an AI token tax.
Artificial intelligence models process information through units called tokens, with businesses using enterprise AI systems frequently paying according to the volume of tokens they consume.
Jensen and Bar Dea argue that the use of AI tokens can be viewed as broadly comparable to paying wages for work performed by machines.
They contend that because the existing tax structure is largely built around taxing human labor, businesses effectively receive an economic incentive to replace employees with machines.
Under their proposal, money collected through an AI token tax could help lower taxes imposed on human labor while also funding a nationwide program giving Americans an ownership stake in artificial intelligence companies.
The plan would have the federal government purchase equity in leading US-based AI companies and then distribute those shares directly to American citizens.
Instead of Washington permanently controlling those investments, individual Americans would hold the shares themselves, allowing them to participate financially in the wealth generated as the artificial intelligence industry expands.
Jensen and Bar Dea are also advocating substantially stronger safety regulations governing advanced artificial intelligence.
They say those safeguards should apply not merely when sophisticated AI models are released to the public, but throughout the development and deployment of the technology.
Their concerns center on the possibility that increasingly independent AI systems could acquire dangerous abilities, operate outside the boundaries established by their creators or be adapted by others for harmful purposes.
Jensen and Bar Dea ultimately argue that Washington does not have unlimited time to decide how it will respond to the transformation already underway.
Their central message is that the United States should continue promoting aggressive AI development while simultaneously making certain that ordinary workers and citizens receive a meaningful share of the resulting economic benefits.
If policymakers fail to strike that balance, they warn, a technological revolution with the potential to generate extraordinary prosperity could instead widen economic inequality, ignite political resistance and introduce dangers that may prove far more difficult to contain once artificial intelligence becomes deeply embedded throughout the economy.
