Jill Lepore Warns: The Artificial State Is Replacing Democracy

In her upcoming book The Rise and Fall of the Artificial State, Pulitzer Prize-winning Harvard historian Jill Lepore delivers an uncomfortable diagnosis: private technology corporations are steadily replacing the functions of democratic government, and almost no one consented to it. Speaking on TechCrunch's Equity podcast, Lepore traced a centuries-long arc from Enlightenment ideals to what she calls "a return to tyranny and mystification in the form of rule by algorithms, corporations, machines."

What Happened

Lepore's central claim is that the "artificial state" is both a real, observable construct and an idea that has been gestating for decades. It's the gradual transfer of governance functions — identity, speech regulation, dispute resolution, even constitutional design — from democratically accountable institutions to private corporations whose leaders operate with little transparency and no electoral mandate.

She points to a cascade of recent examples: Facebook establishing a quasi-judicial "Supreme Court" after the 2016 election, Anthropic hiring a moral philosopher to draft a constitution for its AI systems, and Sam Altman casually floating the idea of an "AI president" on Joe Rogan's podcast. Each incident alone might seem trivial. Together, Lepore argues, they reveal a pattern: Silicon Valley is not just building products — it is constructing a parallel system of governance.

Why It Matters

Lepore's argument lands at a precarious moment. AI systems are embedding themselves into hiring, lending, policing, healthcare triage, and content moderation. These are not market transactions — they are acts of governance. When a model decides who gets a loan, what speech surfaces in a feed, or how a medical claim is prioritized, it is exercising power that used to belong to public institutions with due process obligations.

What makes Lepore's framing distinctive is that she does not blame the technology itself. "I'm not anti-technologist," she said. "I'm married to a computer scientist." Her target is the political philosophy — or lack of one — behind who controls these systems. The 1990s libertarian vision of using the internet to "eradicate the nation-state" has, in her telling, quietly won, not through democratic deliberation but through corporate expansion and regulatory capture.

The Historical Arc

Lepore traces the ideology back to the 1996 Telecommunications Act, shaped by Newt Gingrich's Contract with America, which enshrined a hands-off approach that proved difficult to revisit. But the symbolic origin point, she argues, was Apple's famous "1984" Super Bowl ad. Apple sold the personal computer as an instrument of liberation — the lithe Macintosh smashing the totalitarian mainframe. "That was clever advertising," Lepore said. "I doubt that anybody at Apple really believed the personal computer was going to be an instrument of personal liberation."

Forty years later, the same rhetoric of emancipation — now applied to AI — has curdled into what Lepore calls "a kind of delusional fantasy." The charismatic leaders of today's dominant AI companies invoke science fiction as prophecy while, she contends, misreading even the dystopias they claim to be avoiding.

Builder Impact

For the AI builders and automation engineers who read n8n Lab, Lepore's critique carries practical weight. Every agent deployed, every workflow automated, every decision delegated to a model changes who holds power in a system. The question is not whether AI will reshape governance — it already is. The question is whether the people building these systems understand the political stakes of their work.

Lepore's argument suggests three concrete challenges for builders. First, transparency: if your AI agent makes consequential decisions, can the people affected by those decisions understand how and why? Second, accountability: if your system causes harm, who is responsible — the developer, the platform, the model provider, or no one? Third, consent: did the users and communities affected by your automation ever agree to be governed by it?

These are not abstract academic questions. The AI safety tests themselves are becoming a security risk, as our earlier coverage showed. When AI agents escape testing environments into real-world systems, the line between experimentation and governance blurs irreversibly. The same week Lepore's interview aired, Black Hat USA executives warned that AI agent hacks mark "the start of the most dangerous era in cybersecurity." When corporate AI systems can be weaponized and safety testing frameworks can become attack vectors, Lepore's question — "who consented to this?" — gains uncomfortable traction.

The practical implication for builders is this: governance is not something that happens after you ship. It is embedded in your architecture decisions, your API design, your data handling, and your deployment choices. Every n8n workflow that connects an AI agent to a real-world system — email, calendar, CRM, payment processor — is also constructing a tiny fragment of the artificial state Lepore describes. The responsibility is real, and it compounds with every integration.

What's Next

Lepore's book arrives as Congress debates whether to preempt state AI laws for up to a decade — a proposal that would effectively lock in the current hands-off federal posture while disabling states that have passed their own AI safety rules. The historian's warning is that democratic institutions are racing against corporate power consolidation, and the corporations are winning.

The counter-narrative, which Lepore herself gestures toward, is that the path we're on was not inevitable. The 1990s had fierce debates about what the internet should become. Alternatives existed. The same is true of AI today. The architecture of AI governance — who builds, who controls, who is accountable — is being decided now, and builders are not bystanders to that decision. They are participants whether they choose to be or not.

---

Jill Lepore's The Rise and Fall of the Artificial State publishes later this year. The full TechCrunch Equity podcast interview is available now.