Civic
A new analysis argues that making government data visible is not the same as making government accountable — and the distinction matters enormously as AI enters the picture.
NewsOnScale Staff
July 11, 2026
For decades, the civic technology movement has operated on a foundational assumption: if citizens can see what their governments are doing, power will answer for itself. Open data portals, public records databases, budget visualizers — the tools multiplied, the dashboards proliferated, and reformers declared that sunlight was the best disinfectant. A growing body of research now suggests that assumption is dangerously incomplete, and that the civic tech community has confused visibility with accountability in ways that may have actually helped entrenched interests survive scrutiny.
The argument is not a fringe critique. Researchers and practitioners working at the intersection of anticorruption work and digital governance have been building toward it for years, and it deserves serious attention from anyone thinking about how AI tools now entering the civic space will actually be deployed — and by whom.
## What Transparency Gets Right, and What It Misses
The transparency doctrine rests on a rational-actor model: expose wrongdoing, and affected citizens or watchdog institutions will mobilize to stop it. The problem is that this model quietly assumes a level playing field that rarely exists in practice. A published budget dataset is only useful if someone has the time, expertise, and institutional standing to interpret it, challenge it, and compel a response. In most jurisdictions, that someone is a lobbyist, a well-resourced nonprofit, or a journalist at an outlet with legal backing. Ordinary residents — especially those in communities that bear the sharpest costs of government dysfunction — rarely have those resources.
This is the gap that the transparency-as-accountability framework fails to close. Making data available is a necessary condition for accountability, not a sufficient one. The distance between those two things is where corruption, waste, and institutional inertia have learned to live comfortably.
## The AI Question Nobody Is Asking Loudly Enough
This gap becomes more consequential, not less, as artificial intelligence tools enter civic infrastructure. There is genuine enthusiasm right now — including from well-funded grant programs specifically targeting AI applications in democracy work — about the potential for machine learning to parse procurement records, flag anomalies in public spending, and surface patterns that human reviewers would miss. That enthusiasm is not wrong. These capabilities are real.
But the same structural problem applies. An AI tool that identifies suspicious contracting patterns in a municipal database is performing a transparency function. What happens next — who is notified, what standing they have to act, what institutional mechanisms exist to compel a response — is an accountability function. And most of the current wave of civic AI investment is being directed at the former while the infrastructure for the latter remains underfunded and, in many places, actively weakened.
The risk is that sophisticated transparency tooling becomes a new form of accountability theater: governments can point to open data initiatives and AI-powered dashboards as evidence of good governance while the underlying decision-making remains insulated from meaningful challenge.
## What Accountability Infrastructure Actually Requires
The harder conversation is about power, not data. Genuine accountability requires independent oversight bodies with enforcement authority, legal frameworks that allow affected communities to contest decisions, journalists and civil society organizations with the resources to pursue multi-year investigations, and political cultures that impose real costs on those who obstruct scrutiny.
None of those things are produced by a data portal. None of them are produced by an AI anomaly-detection system, however sophisticated. They require sustained investment in institutions — and in many cases, they require confronting the fact that the governments being asked to become more transparent are the same governments being asked to fund and design the tools that would hold them accountable.
## The Honest Reckoning
Civic technology has produced genuine public goods. The critique here is not that the work was wasted, but that the field has sometimes let the measurable — data published, portals launched, dashboards built — crowd out the harder question of whether any of it changed who holds power and how it is exercised.
As AI accelerates the volume and sophistication of what can be made visible, that question becomes more urgent. The next generation of civic tech investment should be evaluated not on transparency metrics, but on accountability outcomes: Did it help a community stop a harmful decision? Did it produce a consequence for an official who abused public trust? Did it shift power, or just illuminate it?
Sunlight remains necessary. It has never been sufficient.