Why Platform Accountability Fails in Fragile Democracies

Platform accountability is often discussed as if the world has equal regulatory weather. It does not.

In Brussels, platform accountability means risk assessments, researcher access, transparency reports, and regulatory investigations under the Digital Services Act. In Washington, it often means a politically polarised fight over speech, competition, and liability. In London, it means online safety duties and Ofcom enforcement. But in fragile democracies, platform accountability usually means something far more basic: can journalists, civil society groups, election monitors, and targeted communities even get a platform to answer them before harm spills offline?

This is the governance gap that global platform policy still struggles to name.

The problem is not that fragile democracies lack online harms. It is that they often have more severe harms with weaker accountability channels. Electoral manipulation, gendered harassment, religious hate, targeted misinformation, platform-enabled intimidation, arbitrary state takedowns, and internet shutdowns do not occur in a vacuum. They operate in environments where courts may be slow, regulators may be politicised, law enforcement may be selective, and civil society may be treated as a threat rather than a stakeholder.

In such settings, platform accountability fails for at least five reasons.

First, platforms often design enforcement systems around high-resource languages, high-attention markets, and politically powerful jurisdictions. Even when companies claim global coverage, the quality of enforcement depends on language resources, cultural understanding, trusted partner networks, escalation channels, and the commercial importance of the market. A slur, dog whistle, or mobilisation phrase in Urdu, Punjabi, Pashto, Sinhala, Bengali, or Burmese may not map neatly onto the English-language categories inside a global policy team.

Second, fragile democracies often lack reliable public data infrastructure. Researchers struggle to access platform data. Journalists lose access to monitoring tools. API changes break longitudinal work. Meta’s Content Library and API now provide access to public content archives for eligible researchers through an independently reviewed process, but access regimes remain controlled, uneven, and often unavailable to the journalists and civil society actors who are documenting harm on the ground. When CrowdTangle-style access disappears or narrows, the ability to independently scrutinise platform influence also narrows.

Third, state power is not always a neutral regulator. Governments may invoke misinformation, terrorism, blasphemy, public order, or national security to demand removals, block platforms, surveil critics, or criminalise dissent. In Pakistan, human rights groups criticised the 2025 PECA amendments for vague “false and fake information” provisions and broader regulatory powers that could chill online expression. A platform accountability model that assumes the state is always the guardian of rights collapses in contexts where the state is also a source of risk.

Fourth, platforms treat elections in the Global South as episodic crises rather than permanent governance stress tests. Election integrity work often appears during a campaign cycle, peaks around polling day, then disappears. But in fragile democracies, the information environment is shaped long before and long after elections: candidate intimidation, gendered disinformation, religious mobilisation, legal harassment, platform takedowns, surveillance, and shutdowns all affect political participation. UNDP has warned that electoral institutions, especially in Global Majority countries, face capacity gaps, fast-changing technology, and limited engagement with online platforms.

Fifth, accountability still over-relies on transparency reports. Transparency is important, but transparency is not the same as remedy. A platform can publish aggregate numbers while affected users still cannot understand why their content was removed, why abuse was left up, or why escalation failed. Ranking Digital Rights has long argued that transparency is a first step toward accountability, not the endpoint; its 2025 Big Tech evaluation warned that major platforms were stagnating or retreating on key human rights commitments.

The result is a cruel asymmetry. Platforms operate across borders, but accountability is usually jurisdiction-bound. Harm spreads through local social contexts, but moderation systems are built for global scalability. Civil society documents harms in painstaking detail, but escalation channels remain opaque. Governments demand safety, but often define safety as control.

A better model of platform accountability in fragile democracies should start with three shifts.

The first shift is from global policy language to local harm evidence. Platforms should be required to explain how rules are interpreted in local languages and political contexts, especially during elections, communal tensions, and crises. This does not mean outsourcing enforcement to governments. It means building accountable, rights-respecting local expertise.

The second shift is from transparency reports to researcher-grade access. Public dashboards and aggregate reports are not enough. Independent researchers, journalists, and vetted civil society organisations need safe, privacy-preserving, legally protected access to data that allows them to study systemic risks. The DSA’s data-access model is one of the most important experiments in this direction, but it should not become a Europe-only privilege.

The third shift is from content decisions to remedy systems. When a user is wrongly silenced, when a community is targeted, when an election narrative is artificially amplified, or when a platform fails to act on urgent evidence, there must be a traceable path to review and repair. The Oversight Board model has shown that independent review can push policy clarity and enforcement transparency, but it remains platform-specific and limited in scale.

Fragile democracies do not need softer platform accountability. They need stronger, more contextual, and more rights-aware accountability.

The uncomfortable truth is that platform governance built for stable democracies can become dangerously naive elsewhere. If a regulation assumes trusted institutions, but the institutions are captured, it can become a censorship tool. If a moderation policy assumes linguistic neutrality, but harm is coded through local idioms, it can become blind. If a transparency system assumes public data access, but researchers are locked out, it can become theatre.

Platform accountability fails in fragile democracies because it is often built for the world platforms wish existed, not the world users actually live in.