Civic
A new survey reveals deep, bipartisan anxiety about how public agencies collect, store, and share personal information — and a strong public appetite for enforceable accountability.
NewsOnScale Staff
September 6, 2026
There is a version of the civic technology story that gets told at conferences — the one where open data portals and algorithmic transparency dashboards are slowly but surely closing the gap between government power and public understanding. Then there is the version that shows up in survey data, and it is considerably less optimistic.
The Center for Democracy and Technology released findings this week showing that an overwhelming majority of Americans are worried about the personal information held by public agencies — federal, state, and local — and that most want concrete mechanisms for accountability, not just policy statements. The survey did not capture a fringe position. It captured a consensus.
## What the Public Is Actually Worried About
The anxiety is not abstract. Americans are concerned about data collected through benefits systems, law enforcement databases, public health infrastructure, and increasingly, AI-assisted decision-making tools that agencies are quietly deploying to determine everything from child welfare interventions to parole eligibility.
What makes the CDT findings particularly significant for anyone covering the AI agent economy is the timing. Government agencies across the United States are in an active procurement cycle for AI systems. Many of those systems ingest personal data — often data that citizens had no meaningful choice about sharing, because it was tied to a public service they needed. The survey suggests that the public has already reached a conclusion that policymakers have been slow to act on: that passive data collection without genuine accountability is not a neutral act.
The worry is also bipartisan in a political environment where almost nothing else is. That breadth matters. It means the concern cannot be easily dismissed as partisan opposition to a particular administration's technology agenda.
## The Accountability Gap
What the public says it wants — according to the CDT data — is not just disclosure but enforceability. There is an important distinction there. Transparency requirements have been the dominant reform tool in civic technology for the better part of a decade. Publish the data. Post the algorithm. Release the audit. The underlying assumption is that sunlight is sufficient.
It frequently is not. As the Ash Center has argued in parallel research published this week, transparency without consequence tends to produce compliance theater rather than behavioral change. Agencies publish what they are required to publish, and the information sits in formats that are inaccessible to most people and unchallengeable in any practical sense.
Americans appear to understand this intuitively. Wanting accountability is different from wanting a PDF. It implies a mechanism — a right to contest, a path to remedy, a consequence for misuse.
## Why This Matters for the AI Agent Economy
NewsOnScale covers the AI agent economy because the infrastructure being built right now — the pipelines, the decision models, the data procurement practices — will shape what is possible and what is contestable for years. Government is not a passive observer of that build-out. It is an active participant, and in many cases, an early adopter.
When a public agency deploys an AI agent to process benefits claims or flag individuals for additional scrutiny, it is making consequential decisions at scale using data that citizens cannot inspect, through logic that is rarely disclosed, with appeal mechanisms that were designed for a pre-algorithmic era. The CDT survey suggests the public already senses this misalignment, even if they would not use that exact language to describe it.
## What Comes Next
The survey is a diagnostic, not a prescription. But diagnostics matter when the political will to act is being contested. Legislators considering federal data privacy frameworks, agency CIOs making AI procurement decisions, and civil society organizations designing accountability tools all benefit from knowing that public concern here is deep and durable, not episodic.
The harder question — one the data cannot answer — is whether that public concern will translate into institutional pressure before the AI infrastructure inside government becomes too entrenched to meaningfully reform. Based on the current pace of deployment versus the current pace of oversight, that window is narrowing.