How Rwanda Uses AI to Attack Congo: Inside the Digital Influence Machine

Imagine scrolling through X and seeing dozens of accounts praising Paul Kagame, attacking Congo's President Tshisekedi, and pushing the hashtag #ThanksPK — all posted within minutes of each other, all using eerily similar language. Now imagine discovering that every single one of those posts was written by ChatGPT.

That is not a hypothetical. It happened. And researchers caught it red-handed.

The Smoking Gun

In June 2024, researchers at Clemson University's Media Forensics Hub published a study that exposed what they called "Old Despots New Tricks" — an AI-empowered influence operation run by supporters of Rwanda's ruling party, the RPF (Rwanda Patriotic Front), coordinated across X, Facebook, and multiple online platforms.

The evidence was damning. The operatives had made a critical mistake: they forgot to delete the ChatGPT prompts from their posts.

Screenshots captured by the researchers show accounts posting messages that begin with "ChatGPT: Sure, here are 50 content ideas for thanking RPF Inkotanyi, with the hashtag #ThanksPK at the end of each." The AI's response — raw, unedited — was copy-pasted directly onto the platform for the world to see.

This was not a sophisticated operation hiding behind layers of obfuscation. It was a government-aligned campaign using the world's most popular AI chatbot as a propaganda factory, and doing it so carelessly that the machinery was visible to anyone who looked.

What the Campaign Targeted

The messaging revolved around three themes, according to Clemson's analysis:

1. Attacking the Democratic Republic of Congo. AI-generated content depicted President Félix Tshisekedi as "the culprit of the region," producing manipulated images showing him being crucified and caricatured. The campaign amplified narratives that shift blame for the eastern Congo crisis away from Rwanda — the country whose M23 proxy militia, backed by Rwandan troops, has displaced over 7 million Congolese civilians.

2. Promoting Kagame's leadership. Hundreds of AI-generated posts praised Rwanda's economic development, its tourism campaigns, and its hosting of international sporting events. The "Visit Rwanda" campaign — which sponsors Arsenal FC and the Paris-Saint Germain football club — served as one layer of a broader reputation management machine.

3. Drowning out dissent. The researchers found that the objective was not just to spread favorable content, but to dominate the conversation entirely — to ensure that unfavorable messaging about Rwanda's government was buried under a flood of AI-generated noise. Content moderation evasion was built into the strategy: AI tools helped operatives re-version content rapidly so that flagged posts could be replaced before platforms caught up.

The Scale of the Operation

This was not a handful of troll accounts. The Clemson researchers identified it as an extension of a vast network of pro-government social media accounts that Forbidden Stories, the international journalist network, had previously documented.

Forbidden Stories' investigation found "one prominent layer of this reputation management machine involves the use of a vast network of pro-government social media accounts, which have been linked to efforts to promote the state as well as campaigns to harass journalists who have attempted to investigate the government's extrajudicial actions."

AI was the force multiplier. Where human operatives could produce a few posts per day, ChatGPT allowed the campaign to generate content "at scale and speed, in multiple languages and across multiple online platforms," according to the Clemson analysis. The AI also helped re-version content to evade detection — the same message rewritten fifty different ways, each slightly different, each harder for automated systems to flag.

The Congo Connection

For the Democratic Republic of Congo, this campaign is not an abstract concern about online misinformation. It is a weapon in a war that has killed over 6 million people.

While Rwanda's AI-powered accounts flood social media with pro-Kagame content and anti-Tshisekedi propaganda, Rwandan troops continue to back the M23 militia in eastern Congo. A United Nations Group of Experts report documented that between 3,000 and 4,000 Rwandan soldiers are deployed inside Congolese territory, supporting an armed group that the same UN experts say has committed systematic abuses against civilians.

The AI campaign serves a strategic purpose: by controlling the narrative online, Rwanda aims to muddy the waters around its military intervention. If enough content paints Tshisekedi as the problem and Kagame as the visionary, the reality of Rwandan aggression in Congo becomes harder to see through the noise.

The AI-generated caricatures of Tshisekedi — depicting the Congolese president as incompetent or complicit — are part of a dehumanization strategy that mirrors propaganda campaigns used throughout history to justify violence against a target population.

The Freedom House Assessment

Freedom House, in its 2024 report on Rwanda, characterizes the RPF government as one that has "suppressed political dissent through pervasive surveillance, intimidation, torture and rendition of or suspected assassinations of exiled dissidents."

The AI-driven influence campaign, according to the KAS study, appears to be one of several countermeasures aimed at managing dissent — extending the government's physical repression into the digital space.

What It Means

Rwanda's AI campaign is a warning. It demonstrates that AI tools designed for productivity and creativity can be weaponized for political manipulation by any government with a few hundred dollars and a willingness to deceive.

The cost of running a full AI-powered disinformation campaign has collapsed. According to research cited by the Konrad Adenauer Stiftung, a journalist demonstrated that a fully automated fake news website could be operated for just 105 US dollars per year using ChatGPT. Rwanda's campaign shows that the same tools can be turned to coordinated influence operations across major social media platforms.

For the Congolese people — already suffering under the weight of a foreign-backed militia war — this digital offensive represents a second front. The physical violence displaces their bodies. The AI campaign erases their truth.

---

This investigation draws on research by Clemson University Media Forensics Hub, Forbidden Stories, Freedom House, the UN Group of Experts on the DRC, and the Konrad Adenauer Stiftung's 2025 study on AI-Generated Disinformation in Europe and Africa.

Revvmark is an independent investigation platform. Read more at revvmark.com

Based on reporting by REVVMARK