AI Leading Capitalism to 'Extinction': A Political Technology Perspective

An essay pointing out that the rapid development of AI technology could fundamentally shake humanity’s economic system itself is drawing attention. Titled “Extinction-level capitalism,” published on his own site by Matthew Butterick, a lawyer and programmer, the piece depicts the risks posed by AI from a perspective entirely different from conventional narratives of “malevolent AI” or “uncontrollable AI.”

Butterick is the person who, in 2022, upon learning that his own work was included in generative AI training data, filed the first lawsuit questioning its legality. He currently serves as co-counsel for plaintiffs in multiple AI-related lawsuits. In this article, he directly challenges the myth of technological “neutrality.”

The “Emergence” Shown by the Grand Canyon

Butterick begins his argument with a geological metaphor. The Colorado Plateau, where formation began about two billion years ago. Uplift from crustal movements 50 million years ago. A river that began to flow five million years ago. And today’s Grand Canyon. No human intention, technology, or coordination was needed for its formation. The river simply behaved as water, flowing down. The bedrock happened to be there. Butterick defines this as an “emergent effect.”

Common discussions of AI risk focus on cases where AI is misused by “malicious actors” or where AI “malfunctions.” But Butterick’s argument is more fundamental. Like the river that carved the Grand Canyon, AI merely amplifies existing trends—especially capital concentration—leading to irreversible political and economic transformation. No malice or malfunction is necessary.

The Political Nature Inherent in Technology

The theoretical backbone of this essay is the influential work of political theorist Langdon Winner, specifically his 1980 essay “Do Artifacts Have Politics?” Winner argues that technology is not “a neutral tool that can be used well or poorly between good and evil.” He presents two pathways by which technology influences the environment. One is technology intentionally designed to produce specific political effects (e.g., China’s Great Firewall). The other is technology that, regardless of the designer’s intent, reinforces existing social and economic power structures.

Butterick categorizes AI into the latter. AI itself does not possess political will. However, it functions to amplify the existing power structures—particularly capital concentration—of the entities that introduce and operate it. As a result, it gradually but steadily erodes the political system of liberal democracy.

Accelerator of Capital Concentration

Butterick’s greatest concern is that AI accelerates capital concentration. Developing and operating AI technology requires enormous computational resources and data, which overwhelmingly favor large corporations and wealthy nations. The benefits of AI are disproportionately distributed to those who can own and operate it. This movement further accelerates the already ongoing market oligopolization and widening income inequality.

In his essay, he emphasizes that this process requires no “malice” or “malfunction.” AI, by its design, operates to maximize efficiency and productivity. However, the fact that this “efficiency” acts to promote capital concentration is also a political choice.

Limits of the Legal Framework

Butterick, who himself filed AI-related lawsuits, points out that current legal frameworks cannot address this problem. Existing legal systems, such as copyright law and privacy law, do not anticipate the structural transformation caused by AI. Litigation strategies are merely “symptomatic treatments” and do not provide fundamental solutions.

His stance is not that AI should be “banned.” Rather, he argues that after acknowledging the political nature of AI technology, a framework for democratic control should be established. Political debate and citizen participation must be incorporated from the design stage of the technology.

Editorial Opinion

This essay offers a provocative take on existing discussions by framing AI risk not as a sci-fi scenario of “malevolent AI” or “uncontrollable AI,” but as an inherent dynamic of the capitalist system. The perspective of doubting the “neutrality” of technology is not new, but applying it to the real technology of AI is significant.

In the short term, the impact of this article is unlikely to lead directly to policy changes. However, it could serve as an opportunity for the fundamental question of “Is technology neutral?” to come to the forefront in discussions on AI regulation. In the EU’s AI Act and U.S. regulatory proposals, it is time for debates that go beyond mere “safety” or “privacy.”

From a long-term perspective, Butterick’s questioning provides a framework for reconsidering the relationship between AI and capitalism. The possibility that capital concentration accelerated by AI could undermine the institutional foundations of democracy is a point that cannot be ignored. However, the “solutions” he presents lack specificity. Even if we acknowledge the “political nature” of AI, what kind of institutional design can realize its democratic control? Can an international technology like AI be controlled by the democracy of a single nation? Further discussion is needed on these points.

As a question from the editorial team, we ask readers to consider the following: How should we confront the risk that the pursuit of “efficiency” in AI technology could weaken democratic decision-making processes? What specific methods exist to incorporate “democratic values” into technology design? Butterick’s essay provides a new perspective on the debate surrounding AI. But answering these questions is the role of engineers, legal experts, and each and every citizen.

References

• Extinction-level capitalism — Matthew Butterick — Published June 13, 2026