AI tax: analyst calls for protecting the labor market

When an economist whose report shook the valuations of technology giants begins talking about a tax on artificial intelligence, the financial world must listen. Alap Shah, co-author of the notorious Citrini Research study that warned of massive social upheaval due to automation, has called on governments to consider introducing a special AI tax—as a mechanism to protect workers whose professions are rapidly disappearing under the pressure of algorithms.

To understand the weight of this statement, we must recall the context. It was precisely the Citrini Research report that became one of the triggers for a wave of stock market sell-offs, unofficially dubbed the "AI panic". Investors, after reading forecasts about massive job cuts resulting from the deployment of language models and automated systems, began taking profits in the technology sector. Shares of companies that seemed, until recently, to be unstoppable beneficiaries of the AI boom tumbled downward. This episode clearly demonstrated one thing: markets are beginning to factor into prices not only the potential of technologies, but also their social cost.

The idea of an AI tax is not new—it has been discussed at various times by Bill Gates, several European politicians, and academic economists. The logic is simple: companies that replace people with machines save on payroll, pension contributions, and social benefits. The state, meanwhile, loses tax revenue and is forced to bear the costs of supporting laid-off workers. An automation tax is designed to close this gap—to redistribute some of the benefits from AI deployment to society. Shah goes further, proposing that such a mechanism be viewed not as a sanction for technological progress, but as an instrument for its managed implementation.

Technically, implementing such a tax presents a serious challenge. How do you measure the "AI share" in a specific business process? Tax computational power, model licenses, or directly job reductions? Each approach carries its own distortions. A tax on server time could hurt startups using the same capacity for medical diagnostics or climate calculations. A headcount reduction tax is easily circumvented through outsourcing or gradual non-renewal of contracts. The European Union, having already adopted the AI Act, so far avoids direct fiscal regulation, limiting itself to transparency requirements and risk assessment. The United States, with its current political configuration, demonstrates even less inclination toward regulatory experiments in this area.

Nevertheless, the signal sent by Shah's statement is more important than the specific mechanism. A person trusted by the investment community to such a degree that his publications can move markets no longer views the social consequences of AI as a remote hypothetical threat. This is an acknowledgment that automation is already changing employment structures right now—in call centers, legal departments, newsrooms, accounting services. McKinsey and Goldman Sachs, in their latest reports, estimate the automation potential at hundreds of millions of jobs worldwide over the next decade. The gap between the speed of technology implementation and the speed of labor market adaptation is becoming one of the principal economic threats of the era.

For users and the broader audience, the practical consequences of this discussion unfold on several levels simultaneously. First, regulatory pressure on AI companies will intensify—which means the pace of new product releases and their cost may change. Second, the companies themselves, foreseeing future limitations, are already beginning to build elements of "automation social responsibility" into their strategies—if only to get ahead of regulators. Third, investors receive a signal: valuation of technology companies must henceforth include not only revenue growth potential, but also regulatory risks linked to labor policy.

Shah's call is neither populism nor technophobia. It is an attempt to reformat the public conversation about AI: to move it from the plane of "opportunities versus threats" to the plane of "how exactly do we want to distribute benefits and costs". Governments have no answer to this question yet. But its mere appearance on the agenda of analysts capable of moving stock indices means one thing: the technology industry will no longer be able to evade it indefinitely.