A Historian Says Silicon Valley Is Writing Fake Constitutions. She Has a Book to Prove It.

Key points

• Historian and Pulitzer Prize winner Jill Lepore has a new book, The Rise and Fall of the Artificial State, examining how tech companies use government-style language to describe their products.
• Lepore's theory connects platforms like Twitter, which once called itself "a town hall in your pocket", to AI lab documents like Anthropic's Claude Character document, which sets out values for its AI system.
• The argument is that this language is not accidental flattery but a sign that tech companies see themselves as something closer to sovereign powers than to ordinary businesses.
• The book has not yet been published; no release date is confirmed in available sources.

Jill Lepore has a question she keeps coming back to. Why do tech companies talk about their products the way the American founders talked about building a nation?

Lepore, a Harvard historian and Pulitzer Prize winner, addresses that question in her forthcoming book, The Rise and Fall of the Artificial State, first noted by TechCrunch AI. The book's title borrows a political science term: an "artificial state" is one whose borders and rules were drawn by outside powers, not by the people living inside them. Lepore uses it to describe something closer to home.

What is she actually arguing?

Tech companies, Lepore says, dress up product decisions in constitutional clothing, and that habit reveals an unsettling ambition.

The examples are easy to spot once you look. Twitter's old marketing described the platform as "a town hall in your pocket", language that frames a private service as a public institution belonging to everyone. Anthropic, the AI safety company behind the Claude family of AI assistants, publishes what it calls a character document for Claude, a written set of principles meant to guide how the AI behaves. The document borrows heavily from the genre of founding charters.

Lepore's point is not that these companies are lying. It's that the language does real work. When a company frames its terms of service as a constitution, or its content rules as a bill of rights, it quietly claims the moral authority that real governments have to earn through elections, laws and accountability.

The sub-argument, which the book's title hints at, involves science fiction. Tech founders read a lot of it, Lepore suggests, but they tend to read it badly. They absorb the triumphant parts, the inventor who builds a new world, and skip the cautionary ones, the world that pushes back.

Why does this matter for ordinary people?

For anyone who uses a social platform, a chatbot, or any app governed by a "community standards" document, Lepore's argument is worth sitting with.

Those documents look like civic guarantees. In practice they are contracts written by a private company, changeable at any time, with no vote required. Understanding that gap is not a reason to panic, but it is a reason to read the terms a little more sceptically.

The book has not yet been published, so its full evidence and conclusions are not available for review. What the available material suggests is a political and cultural argument rather than a legal or technical one.

Common questions

What is Anthropic's Claude character document?

It is a written document published by Anthropic, the AI company, that sets out the values and personality its Claude AI assistant is designed to follow. Anthropic treats it as a kind of guiding charter for the model's behaviour.

Is Lepore saying AI companies are breaking any laws?

No. The argument is cultural and political: that borrowing government language gives private companies an unearned legitimacy, not that anything illegal is happening.

When does the book come out?

No confirmed publication date is available from current sources.