AI's Catastrophic Risk Is Already in the Welfare Office

Public discussion of artificial intelligence is dominated by dystopian fears: superintelligent machines breaking free of human control, wreaking havoc, and threatening our existence. These hypothetical futures capture imaginations and headlines, and they obscure the more immediate risks AI already poses. While attention stays on dramatic, far-off threats, the dangers already built into everyday systems, from workplaces to government, go largely unaddressed. Who benefits from steering the conversation away from the harms AI is causing now?1

Who the distant threat protects

Tech companies are the first to benefit from this narrative. They’re the main drivers of AI’s rapid development and deployment. Keeping the public focused on apocalyptic futures distracts from how they already use AI to control, monitor, and exploit workers and consumers. In the workplace, AI systems increasingly enforce Key Performance Indicators (KPIs) and metrics that dictate productivity and efficiency. These metrics often clash with workers’ well-being, ethics, or satisfaction, yet AI-driven systems coerce workers into hitting goals many of them fundamentally disagree with. The workplace becomes dehumanizing, a place where employees are tools to be optimized rather than autonomous people with their own values and rights.

With attention on future risks, tech companies can present AI as an innovative force for good and sidestep uncomfortable truths about how it’s amplifying unethical practices now.

Governments and regulators push the same shallow conversation. When the public sees AI risks as distant and hypothetical, there’s less pressure to regulate current uses. AI is already deployed in law enforcement, public services, and other areas where surveillance, bias, and the erosion of civil liberties should lead the conversation. Meanwhile governments can keep expanding AI in ways that increase their control over citizens, without the hard conversations about protecting civil liberties, privacy, and ethical governance today.

Investors and the wider capitalist system also benefit from keeping the conversation narrow. AI is a massive profit opportunity, especially in finance, healthcare, and retail, where automation can streamline processes, cut labor costs, and optimize decisions. A conversation that took seriously AI’s effects on labor markets, privacy, and social inequality would challenge the ethics of pursuing profit at the expense of human dignity. The story of AI as a distant existential threat plays those concerns down and lets investors keep fueling the AI boom without ethical scrutiny.

The media sustains the narrative too. Stories about AI-driven apocalypse capture more attention, and so more revenue, than careful accounts of how AI is already changing society. Meanwhile quieter, more immediate harms such as labor exploitation, erosion of privacy, and algorithmic bias fly under the radar.

Technology designed to control

AI in the workplace is a clear case of what Herbert Marcuse, in One-Dimensional Man, called technological rationality—the idea that technology, under capitalism, is designed not to liberate but to control and suppress. AI-driven KPIs force workers into dehumanizing roles so that capital can keep expanding. Technological rationality of this kind ensures conformity, suppresses dissent, and puts efficiency ahead of autonomy.

Modern power also works through subtle mechanisms that shape individual behavior and identity. Michel Foucault’s concepts of biopolitics (power exercised over a population’s bodies, health and life) and governmentality describe those mechanisms. When AI monitors workers, enforces productivity metrics, and guides consumer choices, that’s biopolitical governance of this kind. It makes people complicit in regulating themselves and normalizes efficiency and optimization as core social values.

AI systems fit Gilles Deleuze’s picture of control societies: diffuse, continuous power that constantly monitors, assesses, and adjusts human behavior. AI shapes workplace dynamics and social governance through pervasive adjustment and surveillance, molding human action across every domain of life.

Powerful actors shape the AI narrative itself to their own benefit. Antonio Gramsci argued, in his theory of cultural hegemony, that dominant groups keep control through coercion and by shaping cultural narratives to present their interests as universal. The simplified narrative of AI as a distant threat is a hegemonic strategy.

The focus on AI’s speculative dangers, rather than its present harms, is also a form of what Mark Fisher called capitalist realism: capitalism keeping itself going by closing off alternative futures. It makes dystopian futures easier to imagine than AI used for ethical, human-centered purposes today. The narrative preserves capitalism’s control over AI and treats the spread of AI through society as inevitable rather than as something people can contest and reshape for the public good.2

Underneath all of it is the alienation Karl Marx analyzed. Marx argued that capitalism alienates workers by turning their labor into a commodity, reducing them to instruments of production. AI intensifies this alienation by turning workers into data points to be optimized, evaluated, and controlled, erasing more of their autonomy and humanity.

AI-driven systems trap workers in the meaningless roles David Graeber described in his critique of bureaucratic violence and the absurdity of modern work. AI reinforces these bureaucratic logics, reducing human creativity and ethical decision-making to compliance with algorithmic metrics.

The conversation around AI needs to shift. Its most pressing dangers are already embedded in the systems that run our work, our governance, and our daily lives. Mitigation starts with holding the institutions that deploy AI accountable for what it’s doing now. What’s at stake is whether AI ends up serving human liberation or domination.