Estimated reading time: 4 minutes · Last updated:
Harvard historian and longtime New Yorker writer Jill Lepore contends that artificial intelligence imperils the shared facts and habits that sustain democratic citizenship. In The Rise and Fall of the Artificial State (Liveright, 336 pp), she characterizes the “Artificial State” as a digital infrastructure of predictive algorithms, bots and corporate-controlled platforms that can reduce public life to machine-tested messaging and systematic attention mining. Lepore frames this danger as the United States marks its 250th birthday and highlights a projection that between 2025 and 2030 AI could drive global data-center power use up by as much as 200 percent.
“Between 2025 and 2030, AI was expected to drive a power increase for global data centers of as much as 200 percent. And, given the pace of the race, the power used by data centers was far likelier to draw on fossil fuels – especially coal – than the power used by other industries.”
Jill Lepore
Key takeaways
- Author and book: Jill Lepore is the author of The Rise and Fall of the Artificial State, reviewed on Sept. 10, 2026; the book is published by Liveright and runs 336 pages.
- Core warning: Lepore coins the term 'Artificial State' for digital systems that can subvert representative government and an informed citizenry through algorithmic manipulation and automation.
- Energy and climate risk: Lepore cites a projection that between 2025 and 2030 AI could push global data-center power demand up by as much as 200 percent and that much of that load may draw on fossil fuels.
Table of contents
Why Lepore sees an 'Artificial State' as a civic threat
Lepore builds her argument from history. She draws a line from earlier worries about mass bureaucracy to today’s digital platforms, arguing that new tools for targeting and automation change how politics is made. The label she gives this system — the Artificial State — bundles technologies and tactics that organise political behaviour through targeted messaging, bots that impersonate people, predictive algorithms and corporate-controlled communications architecture.
That construction is not merely technical. Lepore contends it alters what citizenship means: voting, conversation and civic trust become functions of attention markets and machine testing rather than public argument and deliberation. She traces this dynamic through anecdotes and archival detail, using examples that range from invention stories to modern social media tactics to show how communication infrastructure shapes political possibility.
Where the harms come from — attention, automation and energy
The book separates the technical sources of harm into roughly three buckets: attention-mining algorithms that optimise for engagement, automated agents that can flood public fora, and concentrated platform control that makes large-scale message testing routine. Lepore does not treat these as abstract risks; she ties them to concrete effects on representative institutions and on the everyday capacity of citizens to inspect facts and reason together.
She also raises a non-obvious harm: environmental. As Lepore writes, “Between 2025 and 2030, AI was expected to drive a power increase for global data centers of as much as 200 percent. And, given the pace of the race, the power used by data centers was far likelier to draw on fossil fuels – especially coal – than the power used by other industries.” That linkage makes the governance question wider: managing AI’s civic effects also intersects with energy policy and climate choices.
What Lepore urges and what she leaves open
Lepore’s concluding prescription is cultural and civic rather than legislative. She presses citizens and institutions to revive practices she believes sustain public-minded judgment: inspection of facts, the exercise of reason, and sustained engagement with philosophy, history, literature and the arts. That list is a deliberate refusal of a single tech-policy fix; Lepore presents civic cultivation as an essential bulwark against mechanised persuasion.
At the same time the book is sketchy on precise policy measures. Where she names mechanisms of harm, she rarely specifies the statutes, regulatory designs or procurement rules that governments should pursue. That gap is an opening for further work: researchers and policymakers need empirical studies that measure how algorithmic targeting changes participation, and pilots that test transparency, rate-limiting, public-interest compute procurement, or platform interoperability as interventions.
| Item | How Lepore frames it | Representative example |
|---|---|---|
| Artificial State | A digital communications infrastructure that automates political behaviour and public discourse | Targeted ads, bots, predictive algorithms |
| Energy risk | AI-driven rise in data-center power demand | Projection: up to 200 percent between 2025 and 2030 |
| Civic remedies | Civic and cultural renewal rather than a single technical fix | Reading, study, factual inspection |
Outlook — two ways this could go
The case for
- If civic education and public-media investment rise, citizens may rebuild habits of verification and deliberation that reduce the sway of attention-driven content.
- Technical fixes such as stronger algorithmic transparency, independent audits and purchase of low-carbon public compute could limit both manipulation and the climate footprint Lepore warns about.
The case against
- If platform incentives remain aligned to engagement and targeted persuasion, political discourse may continue to fragment and lose shared factual grounding.
- If data-center energy demand grows toward the projection Lepore cites without a rapid shift to low-carbon power, the environmental cost of AI deployment could deepen social and political strains.
What to be careful about
- Opacity of algorithms and proprietary platforms that prevent independent verification.
- Concentration of communication infrastructure in commercial hands that can prioritise engagement over public interest.
- Rising data-center energy use tied to AI workloads increasing reliance on fossil-fuel generation.
- A policy gap: civic and cultural remedies without concrete regulatory or procurement responses may be insufficient.
The bottom line
Lepore’s book reframes worries about Big Tech by making the civic cost central: when communication systems are tuned to attention and optimisation, the shared practices that enable collective judgment fray. Her solution is not primarily technical; she presses a cultural response — renewed habits of verification, reading and reason — while also pointing to systemic links, such as data-center energy demand, that require policy attention. The test for scholars and policymakers is to translate that cultural verdict into measurable reforms: transparency obligations, empirical studies of influence, and energy policies that prevent AI’s growth from worsening climate and civic harms.
What to watch
- By 2030, compare national or international data-center energy surveys against Lepore’s citation that AI could raise demand by as much as 200 percent between 2025 and 2030.
- Watch for legislative or agency proposals that require algorithmic transparency; no specific date is given in the review.
Frequently asked questions
What does Jill Lepore mean by the 'Artificial State'?
Lepore uses 'Artificial State' to describe a digital communications infrastructure — including targeted messaging, bots and predictive algorithms — that can organise and automate political behaviour and public discourse, potentially undermining an informed citizenry.
How large is the energy risk Lepore highlights?
She cites a projection that between 2025 and 2030 AI could drive global data-center power demand up by as much as 200 percent and warns much of that added load may rely on fossil fuels.
Does Lepore offer concrete policy fixes?
The review says Lepore is vague on specific statutes; her conclusion stresses civic practices — inspection of facts, study of history and literature — rather than prescribing particular regulatory designs.
Related reading
