By Mathias-Felipe de-Lima-Santos, University of New South Wales; Silvia Montaña-Niño, University of Melbourne; Daniel Angus, Queensland University of Technology; and T J Thomson, RMIT.
From being custodians of public knowledge, governments are turning to architects of manufactured ignorance. Amid disappearing evidence, citizens are struggling to hold power to account.
Around the world, governments are quietly deleting, manipulating, or withholding public data at an unprecedented scale, which is a direct threat to democratic accountability.
As official information becomes harder to access, journalists struggle to hold the powerful to account, watchdog agencies lose track of policy failures, and citizens are left increasingly in the dark.
This trend stretches far beyond countries with fragile or emerging democracies. Authoritarian governments such as Venezuela, Indonesia, or the Philippines have long treated data as a political weapon. Similar practices are now surfacing in nations with long-standing democratic institutions such as Australia, Italy, and the United States.
In the United States, the Trump administration removed thousands of public datasets from federal websites—many relating to climate change, environmental monitoring, and federal misconduct. In India, the Modi government has faced allegations of suppressing or altering statistics on issues including extreme poverty and environmental concerns. Meanwhile, Argentina’s Decree 780/2024 places new restrictions on public information access and dismantles long-standing protections for journalistic anonymity, raising alarm among press freedom advocates.
Australia is not immune. Analysts have documented a steady decline in government compliance with Freedom of Information (FOI) requests, alongside significant expansion in Public Interest Immunity claims that allow officials to legally shield documents from scrutiny. Together, these shifts point to growing institutional resistance to transparency, even though digital tools should, in theory, make public records easier than ever to share.
The problem goes beyond access. In many countries, the way data is collected or categorised can distort the reality they claim to measure. Crime statistics have been repeatedly reshaped by governments to influence public perception. In Queensland, Australia, for example, changes to the classification of “minor” versus “serious” offences under the Liberal-National Party coalition government created a new statistical baseline that made crime trends appear more favourable.
Such tactics complicate comparisons over time and leave journalists and researchers struggling to understand what is actually happening on the ground.
When information disappears, is delayed, or becomes prohibitively expensive to obtain, newsrooms face mounting barriers. FOI processes increasingly function as a form of pre-publication control known as SLAPP orders that lead to drawn-out legal delays, exhaustive appeals, and rising processing fees to deter investigations.
In Australia, the backlog for FOI reviews now averages around 16 months. This is long enough for an investigation or story to lose momentum — or for the political conditions to shift entirely, therefore, preventing publication.
An artificial ‘information scarcity’
These are not isolated incidents but signs of a global movement toward engineering “information scarcity”. Governments are moving from being custodians of public knowledge to architects of what Dan Agin calls “manufactured (public) ignorance.” Whether through the erasure of climate records or USAID transparency in the United States, restricted transparency in India and Argentina, or record-low Senate order compliance and FOI refusals in Australia, the result is the same: diminished access to data affects accountability and weakens democratic oversight.
And the stakes are now even higher. As governments around the world turn to artificial intelligence to inform policy, automate services, and shape public decisions, the quality and independence of training data become crucial.
When official datasets are incomplete, obscured, or politically manipulated, the AI systems built on them risk replicating and magnifying those distortions. In this moment of rapid technological adoption, shrinking access to reliable public data is more than a transparency issue; it is a structural threat to the integrity of the future of our society.
As governments dismantle their own information systems, democracies face an uncomfortable question: how can citizens meaningfully hold power to account when the evidence itself is disappearing?
A stark example of resistance to this democratic backsliding comes from Brazil. During the height of the COVID-19 pandemic, the Bolsonaro administration halted the publication of key national health statistics, including infection rates, death counts, and hospital capacity data. The move triggered widespread concern among scientists, journalists, and international health agencies, who warned that suppressing pandemic data in the middle of a public health emergency could cost lives.
In response, a group of volunteer programmers and journalists launched Brasil.IO, a crowdsourcing platform that collected data and reconstructed COVID-19 figures from state health secretariats. Their work became the country’s most reliable source of pandemic data, used by newsrooms, researchers, and even local governments.
Brazil’s experience highlights what happens when official transparency collapses: civil society is forced to rebuild the public record from scratch, often under intense political pressure, simply to ensure citizens have access to life-saving information.
Originally published under Creative Commons by 360info™.
*) DISCLAIMER
Articles published in the “Your Views & Stories” section of en.tempo.co website are personal opinions written by third parties, and cannot be related or attributed to en.tempo.co’s official stance.













































:strip_icc():format(jpeg)/kly-media-production/medias/5243277/original/092151900_1749100247-front-view-cute-little-boy-listening-music.jpg)




