Focusing solely on military-style AI power grabs is too narrow. Extreme power concentration is more likely to emerge from a messy interplay of three factors: active seizures of control, massive economic shifts from automation, and the erosion of society's ability to understand reality (epistemics).

The principle that a small group will always emerge to lead is a fundamental law of human organization. This isn't limited to geopolitics or massive corporations; it's a fractal pattern observable in every group, including one's own family.

Extreme wealth inequality creates a fundamental risk beyond social unrest. When the most powerful citizens extricate themselves from public systems—schools, security, healthcare, transport—they lose empathy and any incentive to invest in the nation's core infrastructure. This decay of shared experience and investment leads to societal fragility.

The concentration of wealth where the top 10-20% capture 70-80% of the economic pie is fundamentally unstable in a democracy where everyone gets a vote. This economic reality serves as a political invitation for populist demagogues, making the rise of radical socialist ideas a predictable and dangerous outcome.

Robert Solow posits that rising inequality isn't just an economic issue; it's a political one. Initial economic disparities lead to political inequality, which then allows the powerful to shape laws (like deregulation) in their favor, further concentrating wealth and reinforcing the initial inequality.

While a fast AI takeoff accelerates some risks, slower, more gradual AI progress still enables dangerous power concentration. Scenarios like a head of state subverting government AIs for personal loyalty or gradual economic disenfranchisement do not depend on a single company achieving a sudden, massive capability lead.

Don't expect corporate America to be a bulwark for democracy. The vast and growing wealth gap creates an overwhelming incentive for CEOs to align with authoritarians who offer a direct path to personal enrichment through cronyism, overriding any commitment to democratic principles.

Meredith Whittaker argues the biggest AI threat is not a sci-fi apocalypse, but the consolidation of power. AI's core requirements—massive data, computing infrastructure, and distribution channels—are controlled by a handful of established tech giants, further entrenching their dominance.

Basic efficiency—doing things in bulk is cheaper—drives the growth of massive index and private equity funds. Harvard's John Coates argues this economic good creates a political problem, as the resulting concentration of influence in a few firms is at odds with the democratic principle of dispersed power.