Social media has a problem; anyone who has spent more than twenty minutes scrolling through any platform in 2026 already knows this.

Misinformation travels faster than corrections. Coordinated disinformation campaigns have influenced individual actions, elections, derailed public health responses, and incited real-world violence.

Governments across the world are under genuine pressure to do something about the information environment, and most of them are struggling to figure out what that something should look like without becoming the very threat they claim to be fighting.

Kenya is now the latest country to make a move. The Kenyan government is seeking parliamentary approval for a KES 2.7 billion allocation, approximately $20.8 million, to deploy artificial intelligence tools that would monitor social media, track public sentiment around government policy, and establish a centralised communications hub for coordinating government messaging.

The State Department for Broadcasting and Telecommunications has justified the request by pointing to the spread of disinformation, misinformation, and what it calls "malinformation" on social media platforms.

On paper, it sounds like responsible governance. In practice, the picture is considerably more complicated than the average eye can see.

The Justification Is Real. So Is the Risk.

It would be intellectually dishonest to dismiss Kenya's concerns as manufactured. The country has a documented history of online disinformation affecting its political landscape, from coordinated harassment campaigns targeting journalists to the spread of false information during election cycles.

Social media platforms, largely headquartered thousands of miles away in California, have repeatedly demonstrated that they will not self-regulate at the pace or scale that countries in the Global South need them to.

When platforms fail to act, and governments lack the tools to respond, ordinary citizens absorb the consequences.

The budget breakdown is revealing. KES 400 million goes toward AI software to analyse social media sentiment. KES 926 million funds the establishment of a National Communication Centre. Another KES 242.79 million covers media and customer relations management software.

The remaining allocations address the Kenya News Agency infrastructure and ministry operations. The stated purpose of the National Communication Centre is to package and distribute government information through coordinated messaging, a development that is being framed as a solution to the problem of government ministries sending contradictory public communications.

That particular justification is reasonable on its face. Governments do struggle with internal coordination, and mixed messaging during crises has real consequences. But sentiment analysis software is a different category of tool entirely.

It does not just flag false information. It monitors how people feel about government policy in real time, maps the emotional temperature of public discourse, and identifies patterns in opposition and dissent.

The government has not explained, with any precision, what it intends to do with that information once it has it. That gap is where the fear lives, and it is not an unreasonable fear.

Other Countries Tried This. The Results Were Similar.

Kenya is not operating in a vacuum. Several countries have built or deployed AI-powered social media monitoring systems, and the outcomes paint a complicated picture that governments seeking similar tools rarely acknowledge publicly.

China's social credit system and its accompanying internet surveillance infrastructure represent the most extreme end of the spectrum. The apparatus monitors online behaviour, flags dissent, and links digital activity to real-world consequences for citizens.

The result is one of the most comprehensively censored information environments in the world, where criticism of the government is not just discouraged but systematically eliminated.

China is, deliberately, not a model most democratic governments will cite, but it demonstrates what the infrastructure of sentiment monitoring becomes when political will to suppress dissent is present.

Closer to Kenya's democratic context, Singapore has deployed sophisticated online monitoring tools under the Protection from Online Falsehoods and Manipulation Act, which grants the government authority to order corrections or removal of content it designates as false.

The law has been criticised extensively by press freedom organisations for giving the government unchecked power to define what counts as a falsehood. Legitimate political commentary has been caught in its net.

Ethiopia has used digital surveillance infrastructure to track and, in documented cases, prosecute critics and journalists. The Ethiopian government did not frame these systems as censorship tools at the point of deployment. It was framed as public order measures.

The pattern is consistent across different political systems and continents: the stated purpose of social media monitoring infrastructure and its operational use diverge over time, particularly when governments face domestic political pressure.

The tool built to fight misinformation becomes the tool used to manage narratives. This is not a hypothetical, it is a documented trajectory that has played out over the years.

Kenya itself has already demonstrated this dynamic. A Kenyan court, in 2025, ruled against government attempts to block social media access, a move that drew significant attention across the continent.

The current AI monitoring proposal is widely read by critics as a strategic pivot: rather than blocking platforms outright, which courts have rejected, the government builds the capacity to monitor and respond to content instead.

The objective of controlling the online narrative seems not to be changing, but the method by which the intent is to be executed is changing.

What This Means for Free Speech and Why It Matters

The fundamental tension in this debate is not between misinformation and truth. It is between state capacity and civil liberty.

Every government that has ever restricted speech has done so under the banner of protecting something, public order, national security, or social cohesion. The banner is almost always legitimate, but the mechanism is almost always the problem.

Kenya's proposal deserves scrutiny, not because fighting disinformation is wrong, but because the architecture being built extends well beyond that stated goal. Sentiment analysis of government policy discussions is not a fact-checking tool. It is a political intelligence tool.

The National Communication Authority, designed to coordinate government messaging across ministries, combined with real-time sentiment monitoring, creates a feedback loop where the government tracks public reaction to its communications and adjusts its messaging accordingly.

In the hands of a government committed to transparency and accountability, that is sophisticated public relations. In the hands of one more interested in control, it is something closer to a propaganda infrastructure.

The KBC salary crisis, sitting alongside this budget request, says something important about priorities. Kenya's public broadcaster cannot cover legally mandated salary deductions for its staff, operating with a KES 592 million deficit in its payroll budget.

The country's existing public information infrastructure is underfunded and deteriorating. But KES 2.7 billion is available for AI surveillance and communications management. That is a choice, and choices reflect values.

Civil society in Kenya, press freedom organisations, and opposition voices are right to demand specific, legally binding answers to specific questions before this funding is approved.

What legal framework governs how collected sentiment data is stored and used? Who has access to it? What oversight mechanism exists outside the executive branch? What are the explicit prohibitions on using this system to target journalists, activists, or political opponents?

Social media does need to be addressed. The disinformation problem is real, and its consequences are serious. But the solution cannot be a system that gives any government unlimited visibility into what its citizens think and feel about it, with no independent accountability for how that visibility is used.

Kenya is at a decision point, the infrastructure it builds now will outlast the administration that builds it.

That is the argument every government deploying surveillance tools counts on people forgetting, and the argument citizens and legislators cannot afford to let go of.