Civic
A new analysis from Harvard's Ash Center argues that transparency tools have become a substitute for accountability rather than a path toward it.
NewsOnScale Staff
October 9, 2026
There is a seductive logic to transparency technology. Build the dashboard, publish the contracts, release the budget data in machine-readable format — and democracy, presumably, does the rest. It is a logic that has attracted hundreds of millions of dollars in philanthropic funding, spawned thousands of civic-tech startups, and reshaped how reformers in dozens of countries think about fighting corruption.
A new analysis from Harvard Kennedy School's Ash Center for Democratic Governance and Innovation is now challenging that logic with methodical force. The paper, drawing on years of comparative case studies across civic-tech deployments in the Global South and established democracies alike, reaches a conclusion that should unsettle anyone who has ever celebrated a government data portal launch as a win for accountability: transparency, on its own, does not reliably reduce corruption. In many documented cases, it does not reduce it at all.
## The Gap Between Data and Action
The Ash Center's core finding is structural. Transparency tools assume the existence of actors who are both willing and empowered to act on disclosed information — investigative journalists with legal protection, opposition parties with genuine political standing, courts insulated from executive pressure, or civil society organizations with the capacity to translate raw procurement data into enforceable complaints. Where those actors exist and function, transparency tools can amplify their effectiveness. Where they do not, open data portals become sophisticated theater.
This is not a minor qualification. It describes the political reality of a large share of the countries where civic-tech investment has been most enthusiastic. Publishing sanitized budget data in an environment where the press is concentrated, the judiciary is compromised, and civil society organizations face registration barriers or funding restrictions does not create accountability. It creates the appearance of accountability — which, for governments inclined toward performance over substance, may be precisely the point.
The analysis also identifies what it calls the "participation fallacy": the assumption that making data available is equivalent to making it usable. Most disclosed government datasets require significant technical capacity to parse, contextualize, and verify. The populations most harmed by corruption — rural communities, informal workers, low-income households — are least positioned to engage with spreadsheets and API endpoints. The civic-tech tools built to serve them frequently serve, in practice, a narrow layer of urban, educated intermediaries whose interests and priorities may not align with those of the communities nominally being helped.
## What Actually Works
The paper stops short of condemning transparency initiatives wholesale, and that restraint is earned. The cases where open data interventions have produced measurable accountability outcomes share identifiable features. They involve not just disclosure but feedback mechanisms — ways for citizens to submit complaints, track responses, and escalate unresolved issues through multiple channels. They are embedded in coalitions that include legal aid organizations, unions, or local government ombudsmen with real enforcement capacity. And they are sustained over years, not launched as pilot programs and quietly defunded when grant cycles end.
The distinction matters enormously for how platforms, foundations, and governments allocate resources in the AI era. The current wave of civic AI tools — automated procurement anomaly detection, natural-language interfaces for government databases, AI-assisted analysis of public spending — inherits all of the assumptions of the transparency paradigm, and then layers on new ones about computational power as a substitute for political will.
## The Harder Question
For an AI agent economy increasingly pitching automated governance tools to public institutions, the Ash Center's analysis is a direct challenge. Building better tools for surfacing corruption signals is genuinely valuable work. But the bottleneck in most accountability failures is not analytical — it is political and institutional. No model, however capable, compels a prosecutor to file charges, protects a whistleblower from retaliation, or prevents a regulatory body from being captured by the interests it oversees.
The civic-tech sector has spent a decade building instruments of extraordinary precision and deploying them into environments that lack the institutional infrastructure to act on what those instruments find. The next decade will require an honest reckoning with that gap — and a willingness to fund the harder, slower, less demonstrable work of building the accountability systems that make transparency meaningful.