Americans Don't Trust Government With Their Data — And the Numbers Back Them Up

NewsOnScale Staff

September 18, 2026

Something rare is happening in American political life: people who agree on almost nothing else are agreeing on this. According to new research from the Center for Democracy and Technology, concern about personal data held by government agencies is not a partisan flashpoint — it is a shared anxiety cutting across political identity, geography, and income level. That breadth should command attention from anyone watching how AI is being absorbed into public administration.

The CDT survey, released this week, found that large majorities of Americans are worried about what data federal, state, and local agencies collect about them, how securely it is stored, and whether it is being shared with third parties — including law enforcement and private contractors — without meaningful consent or transparency. Respondents expressed particular concern about the lack of accessible information explaining what data is held and how it is used. In other words, the problem is not just exposure. It is invisibility.

## The Opacity Problem

This is precisely the kind of finding that tends to get buried under more theatrical political news, but it maps directly onto one of the most consequential infrastructural shifts of the decade. Government agencies at every level are deploying AI systems — for benefits determination, predictive policing, child welfare screening, tax auditing, and more — and those systems are, by design, data-hungry. They require not just the information citizens voluntarily provide, but aggregated records, behavioral inferences, and third-party data purchases that most people have no awareness of.

The CDT data suggests that citizens have sensed this expansion even without knowing its technical contours. Intuition, in this case, is tracking reality. The Federal Privacy Act of 1974 — the primary statute governing how agencies handle personal records — was written before the internet existed and has been amended only piecemeal since. There is no comprehensive federal data protection law. There is no unified public dashboard where a citizen can request a consolidated account of what the government knows about them. The opacity that survey respondents are registering is not a perception gap. It is an accurate read of the legal and institutional landscape.

## Accountability Without Architecture

What makes this survey particularly relevant to the AI agent economy is the accountability vacuum it describes. When a government agency uses an automated system to deny a disability claim or flag a benefits recipient for fraud review, the data underpinning that decision may have been sourced, processed, and weighted in ways the affected individual cannot examine, challenge, or even identify. The CDT findings suggest Americans are beginning to register this asymmetry — that the government can know a great deal about them while they are permitted to know very little about how that knowledge is used.

The demand side of the survey is equally instructive. Respondents did not simply express fear; they expressed a preference for structural accountability. They want to know what data is collected, they want the right to access and correct it, and they want meaningful consequences when agencies mishandle it. These are not radical asks. They are baseline expectations that most peer democracies have codified into law.

## Why This Moment Matters

The timing of this research is not incidental. Across federal agencies, AI procurement is accelerating. The current administration has moved to streamline AI adoption in government services while simultaneously rolling back certain oversight requirements put in place under previous executive guidance. State and local governments, often with fewer resources and less scrutiny, are adopting off-the-shelf AI tools from vendors whose data practices are governed primarily by contract rather than statute.

In that environment, survey data showing broad public demand for accountability is not just a political story. It is a signal about the gap between where government data infrastructure is heading and where the public expects it to be governed. That gap — between the technical reality of AI-powered public administration and the legal architecture designed to constrain it — is the accountability story of the next decade.

The CDT findings will not, by themselves, produce a federal privacy law or force a single agency to publish its data practices. But they do something important: they establish, with evidence, that this is not a niche concern for civil libertarians. It is a mainstream worry, shared by the people these systems are being built to serve.