It Started in Minneapolis
In January 2026, federal agents came to Minneapolis. The president’s words painted the city as lawless — a place that protected criminals, thugs, and thieves. But that wasn’t what people who lived there saw. What they saw was neighbors defending neighbors — insisting that human dignity matters regardless of immigration status.
The frustration wasn’t just about the policy. It was about the lie embedded in the national narrative. Minneapolis wasn’t protecting criminals. Minneapolis was protecting people. And the people who actually lived there had no trusted, reliable mechanism to say so.
That gap — between what a community actually thinks and what the national conversation claims it thinks — is what Staked Voice is designed to close.
The measurement tools are broken
Democracy depends on a feedback loop: citizens express preferences, representatives respond, citizens evaluate the response. When that loop breaks, institutions lose the signal they need to function, and concentrated interests fill the void.
The bots are winning. In 2026, Nature published research showing AI chatbots infiltrate online surveys at a 99.8% success rate. The underlying PNAS study called it a potential existential threat to social science. If bots can corrupt scientific research, they can corrupt democracy.
The disconnect is structural. Political scientists analyzed 1,779 policy issues and found that average Americans’ preferences had a statistically near-zero impact on policy outcomes. Economic elites and organized interest groups had substantial independent influence. (Gilens & Page, 2014)
The silence is self-reinforcing. Between 40–67% of Americans self-censor their political views — confirmed by a 20-year longitudinal study. This exceeds McCarthyism, when only one in eight Americans felt afraid to speak.
Professional polling — the primary instrument of democratic feedback for nearly a century — is in structural decline. Response rates have fallen below 6%. Gallup ended 88 years of presidential approval polling. And Elisabeth Noelle-Neumann’s Spiral of Silence takes hold: people assume their neighbors disagree when the opposite may be true.
Orwell saw this coming
“We are all drowning in filth. When I talk to anyone or read the writings of anyone who has any axe to grind, I feel that intellectual honesty and balanced judgment have simply disappeared from the face of the earth. Everyone’s thought is forensic. Everyone is simply putting forward a ‘case’ with deliberate suppression of his opponent’s point of view.”— George Orwell, diary entry, 1942 · expanded in “Notes on Nationalism” (1945)
In “Notes on Nationalism”, Orwell identified the mechanism: when shared reality dissolves, thinking becomes purely adversarial — advocacy masquerading as analysis. He defined “nationalism” broadly as the habit of identifying with any single power unit and placing it beyond good and evil. The specifics are interchangeable. The frenzy is the constant.
Modern research confirms what he intuited: political parties have grown more polarized since World War II, while the average voter remains just as moderate on key issues as they always have been. The polarization isn’t bottom-up. It’s driven by elite incentives to differentiate constituencies.
Stanford research documents the structural consequence: moderates “hide their viewpoints through self-censoring and expressing preferences inconsistent with their private beliefs — a phenomenon known as preference falsification.” The measured middle becomes socially invisible, making the public square appear more extreme than it actually is.
The moderate majority doesn’t disappear. It goes silent. And into that silence flows tribal certainty, manufactured consensus, and algorithmic amplification of extremes. The gap isn’t between left and right. It’s between what people actually think and what gets measured.
Research published in Social Science Journal found that feelings of uncertainty drive identification with radical groups — extremism offers cognitive relief by simplifying a confusing world into a legible enemy. Georgia State researchers show that repetitive zero-sum framing creates the atmospheric conditions for this: once opponents are cast as existentially evil, common ground becomes structurally impossible.
The consequences are already here
The problems above aren’t theoretical. They’re operational.
The United States is no longer classified as a stable democracy. The Polity data series now labels it an “anocracy.” International IDEA’s 2025 report confirms the ninth consecutive year of global democratic decline. The Financial Times analyzed 139 historical episodes: US backsliding outpaces the early trajectories of Russia, Turkey, Hungary, and Venezuela combined.
In 2025, democratic threats became operational. The DOJ targeted ActBlue’s 14 million donors. The FBI infiltrated encrypted Signal chats of immigrant-rights activists. Apple and Google removed ICE-tracking apps under federal pressure. Free Press documented approximately 200 attacks on the First Amendment in a single year.
This is an economic problem too
Economists Acemoglu, Naidu, Restrepo, and Robinson published rigorous causal analysis in the Journal of Political Economy: democratization increases GDP per capita by approximately 20% over 25 years. Democratic institutions directly produce economic growth — through rule of law, public goods investment, broad-based demand, and protection for innovation.
Strip away democratic voice, and you strip away the institutional foundation of broad-based prosperity. When 85% of Americans believe their representatives don’t care, and 40–67% are afraid to say what they think, we’re not just losing democracy. We’re losing the infrastructure that makes shared prosperity possible.
AI governance is being decided without you
72% of Americans have concerns about AI’s role in weapons and surveillance. None had a structural role in the most consequential AI governance decisions of 2026. Researchers are running deliberative experiments — but no binding infrastructure yet gives the public a voice in AI decisions before they’re made.
That outcome depended on one company’s willingness to absorb a loss and a national security blacklisting. That’s not governance. That’s courage compensating for the absence of governance. It’s the kind of gap Staked Voice is designed to help close.
The gap Staked Voice is designed to fill
The problem is not that people don’t have political views. It’s that no existing infrastructure can surface those views with three simultaneous properties: resistance to bot manipulation, protection from surveillance, and verifiable geographic granularity.
Traditional polling fails on bot resistance and is declining in reach. Social media fails on all three. Voting happens too infrequently to serve as a continuous feedback mechanism. Staked Voice runs an open caucus process — continuous, privacy-preserving, bot-resistant measurement of what communities actually care about, built on infrastructure that cannot be compelled to disclose individual responses.
The US has terminated 86% of USAID democracy-support awards — approximately $80.5 billion in cancelled democratic infrastructure. Political measurement is collapsing from the inside. The tools exist. The threat is documented. The question is whether anyone builds alternatives before the window closes.
Ready to participate? Join the waitlist →
← Back to Staked Voice