In the sweltering, red-dust streets of Bangui, the capital of the Central African Republic, a quiet but relentless battle for hearts and minds is being fought. It is not waged with bullets or armored vehicles, but with words, images, and the silent logic of artificial intelligence. According to a report by the US-based AI company Anthropic, Russian agents have been deliberately using its chatbot Claude to spread pro-Russian influence, monitor opposition politicians, and manufacture propaganda that serves the Kremlin’s interests in this fragile, conflict-scarred nation. The report, titled “Detecting and Countering Misuse of AI,” published in September, paints a picture of a sophisticated digital operation hiding in plain sight. Much of the content produced with Claude’s help was broadcast by Radio Lengo Songo, a station that the international investigative collective All Eyes on Wagner says is funded by Russia. The collective, a network of journalists, academics, and experts tracking the Wagner Group and its successor structures like the Africa Corps, has long documented Moscow’s creeping influence in the country. Now, with the help of Anthropic’s findings, the world can see how AI has become another weapon in that arsenal—a tool to shape narratives, silence dissent, and rewrite reality for a population already weary of instability.
The mechanics of this operation are both chilling and disturbingly mundane. Anthropic’s report describes how Claude was prompted to generate content that painted the Central African government and the Africa Corps mercenaries in a glowing, heroic light, while casting France and the domestic opposition as treacherous enemies. The AI was also used to monitor opposition figures, gathering data on their activities and, in some cases, framing them as armed militants in an effort to provoke security forces into taking action against them. Anthropic says Claude refused the latter request, but the very attempt reveals the intent. Lou Osborn, a member of All Eyes on Wagner who contributed to the research, sees this as part of a long-standing pattern. He points to the arrests of people who were not necessarily close to President Faustin-Archange Touadera’s regime, arrests that have been happening for years. “The arrests were the subject of online campaigns in newspapers and on the radio which portrayed those individuals either as proxies for France or Western intelligence services, or as people preparing a coup against Touadera,” Osborn told DW. This is not new, he insists; it is the same playbook that Wagner has used across the continent, now supercharged with the speed and scale of artificial intelligence. The report does not name the opposition figures who were monitored, nor does it specify what data was collected or who ultimately received it, leaving many questions unanswered.
For the opposition in the Central African Republic, the report is less a revelation than a confirmation of what they have long suspected. Martin Ziguele, chairman of the Movement for the Liberation of the Central African People, says the government and its Russian allies are determined to get rid of them by any means necessary. He recalls the dark days of 2020, when the CPC, a coalition of rebel groups, attacked Bangui. In the aftermath, he says, several media outlets went out of their way to fabricate links between opposition leaders and the rebels. “Strangely enough, several media outlets did everything in their power to establish alleged links between certain opposition leaders and the CPC,” Ziguele told DW. “There were also attacks against the opposition claiming that we were being paid by foreign countries. There was an attempt to pass a law stating that foreign agents were present in the country. All of this relates to the activities of the Wagner Group, as does this latest revelation.” For Ziguele and his colleagues, the use of AI is merely the latest chapter in a long, exhausting struggle. The stakes are not abstract: in a country where political violence is never far away, being labeled a traitor or a coup plotter can have fatal consequences. The opposition is not just fighting for votes or policy; it is fighting for survival in a political landscape where the lines between truth, propaganda, and outright fabrication have been deliberately blurred.
Russia, as expected, dismisses the entire report as baseless. Dmitri Sytyi, a representative of the Russian mission in the Central African Republic, shrugged off the findings when asked by DW. “Yes, I heard about [the report] in the news,” he said. “But there are no concrete facts in it. These are vague allegations, without precision and without evidence.” It is a familiar response, one that Moscow has deployed time and again in the face of accusations about its activities in Africa. But the report suggests that the Central African Republic is far from an isolated case. According to Anthropic, similar operations were detected in the Democratic Republic of the Congo, Kenya, Sudan, and Mali. In the DRC, the company identified 318 articles published by a network of about 70 fake news sites, with content distributed by more than 250 fictitious accounts on X. These articles supported the Congolese government’s positions in its tense conflict with Rwanda. In Kenya, Claude was allegedly used to generate batches of 50 messages at a time, all supporting the government and attacking the opposition. In Sudan, AI was reportedly used to fabricate witness statements for the United Nations, part of an operation attributed to individuals in the United Arab Emirates. And in Mali, an adviser with ties to intelligence services allegedly used Claude to develop a platform for monitoring telecommunications, designed to target 25 million SIM cards while bypassing the need for judicial authorization. Each case is different, but the underlying pattern is the same: AI is being used to manipulate public opinion, undermine democratic processes, and consolidate power.
Anthropic says it uncovered and shut down all of these operations between December 2025 and August 2026, primarily by deleting the accounts involved and strengthening its detection systems. The company’s willingness to investigate and expose these abuses is commendable, but it also highlights a troubling reality: the same tools that promise to democratize knowledge and creativity are being weaponized by state actors and their proxies to spread disinformation and surveil their enemies. The report is a reminder that AI is not inherently good or evil; it is a mirror of human intentions. In the hands of a journalist, it can help uncover the truth. In the hands of a propagandist, it can bury it under an avalanche of lies. For the people of the Central African Republic, the consequences are already visible. They have lived through coups, rebellions, and foreign interventions. Now they must navigate a world where the voices they hear on the radio or read online may not be human at all, but the carefully crafted output of a machine designed to manipulate them. The fight for the country’s future is no longer just about who controls the guns; it is about who controls the narrative. And in that fight, AI has become the newest and perhaps most insidious weapon of all.
As the international community grapples with the implications of AI-driven influence operations, the case of the Central African Republic serves as a stark warning. It shows how easily technology can be turned against vulnerable populations, how quickly truth can be twisted into propaganda, and how difficult it is to hold anyone accountable. The report by Anthropic is a step in the right direction, but it is only a first step. Governments, tech companies, and civil society must work together to build safeguards that prevent AI from being used to silence dissent, spread lies, and destabilize nations. For the opposition leaders in Bangui, for the journalists trying to report the truth, and for ordinary citizens trying to make sense of a chaotic world, the stakes could not be higher. They are not just fighting for political power; they are fighting for the very idea that truth still matters. In a world where a machine can be programmed to hate, to deceive, and to destroy, the defense of human judgment, human empathy, and human connection has never been more urgent. The story of Claude in the Central African Republic is not just a story about AI; it is a story about us, our vulnerabilities, and our choices. And it is far from over.