The modern developer or product manager rarely wakes up intending to dismantle the state. Most days are spent optimizing latency, refining prompt templates, or debating the nuances of a new API. Yet, there is a creeping shift in how these tools are deployed. We have moved from using AI to summarize documents to using AI to determine what is true, what is ethical, and who is allowed to speak. This transition is happening quietly, embedded in the terms of service and the system prompts of the world's most powerful models, turning technical implementation into a form of silent legislation.

The Mimicry of Sovereignty

Jill Lepore, a Harvard historian and staff writer at the New Yorker, argues that this shift is not a coincidence but a systemic movement toward what she calls the Artificial State. In her upcoming book, The Rise and Fall of the Artificial State, Lepore analyzes how Big Tech companies are rapidly absorbing the functions of democratic governments. She defines this phenomenon as a regression into a form of autocracy and mysticism, where governance is handed over to algorithms, corporations, and machines rather than elected representatives.

This is not a theoretical fear but a documented pattern of mimicry. Lepore points to the 2016 establishment of Facebook's Oversight Board—essentially a corporate Supreme Court—created to handle content moderation disputes that the company could no longer manage through simple policy. More recently, the trend has moved into the very architecture of artificial intelligence. Anthropic has employed moral philosophers to draft a constitution for its models, creating a set of guiding principles that dictate the AI's behavior and ethical boundaries. Even Sam Altman, CEO of OpenAI, has openly discussed the concept of an AI President in podcast appearances, signaling a willingness to imagine a world where machine intelligence assumes the highest office of state.

Lepore traces the psychological roots of this ambition back to 1984. The famous Apple Macintosh commercial depicted the mainframe computer as a symbol of a totalitarian state, promising that the personal computer would be the tool of individual liberation. However, Lepore argues that the marketing rhetoric of the eighties has curdled. The dream of liberation has been replaced by a delusional fantasy among tech leaders who now view the machine not as a tool for the citizen, but as the governor of the citizen.

From Efficiency to Usurpation

To understand how we arrived here, one must look at the trajectory of private sector involvement in public functions. Initially, the movement was driven by pragmatism. Governments outsourced technical tasks to private firms to achieve greater speed, lower costs, and higher efficiency. It was a gradual, almost accidental slide where the state relied on the private sector to modernize its infrastructure. However, Lepore suggests that over the last 20 to 25 years, this dynamic has shifted from passive assistance to an intentional strategy of usurpation.

This shift is rooted in the political climate of the 1990s, specifically the influence of libertarian futurists who envisioned an internet that would render the concept of the nation-state obsolete. This ideological drive found a legislative home in the Telecommunications Act of 1996. The act was deeply influenced by the political currents of the time, including Newt Gingrich's Contract with America, which sought to shrink the role of government and deregulate the digital frontier. This legislative environment provided the structural foundation for the current internet, allowing a few corporate entities to build the digital roads and courts that the public now relies upon.

Consequently, the current competition between AI giants is no longer just about who has the most parameters or the lowest perplexity. It is a race for governance authority. When tech leaders adopt the language of constitutions and supreme courts, they are signaling that they no longer wish to be mere tool providers. They are positioning themselves as the primary arbiters of social consensus and legal judgment. The danger lies in the fact that the public is increasingly trading democratic decision-making power for technical efficiency, accepting the rule of the algorithm because it is faster than the rule of law.

For those building and deploying these systems, the critical metric is shifting from efficiency to legitimacy. When a company implements an AI-driven internal regulation or an automated decision-making system, it is not just optimizing a workflow; it is exercising power. If an AI constitution replaces a democratic consensus, the resulting system may be efficient, but it lacks the fundamental legitimacy required for stable governance. The risk is that the governance gap created by this transition will eventually lead to a crisis of authority that no amount of compute can solve.