AI Firm Reveals Alleged UAE Influence Campaign on Sudan Conflict

Anthropic, the developer of the Claude AI system, has identified and disrupted a large-scale influence campaign that leveraged its technology to shape discourse around the war in Sudan. According to a threat intelligence report released on September 10, the company linked the operation with high confidence to officials within the United Arab Emirates government. The disclosure, reported by Agence France-Presse, outlines a network of approximately 300 inauthentic social media accounts designed to mimic independent voices while advancing specific political agendas.

The alleged operation extended beyond digital manipulation to include the fabrication of evidence for international bodies. Anthropic stated that the network used AI to draft testimonies for submission to the UN Human Rights Council, specifically instructing that the texts omit any reference to the UAE. The company noted that these documents were intended to be presented as independent witness accounts, thereby lending false credibility to a particular narrative regarding the conflict in Sudan.

Fabrication of Independent Testimonies

The report details how the network appropriated the identity of a genuine Swiss organization to disseminate human rights reports that appeared authentic. Furthermore, the operation targeted the UN system directly by preparing full testimonies in the names of two individuals associated with a real Sudanese human rights organization. The goal, according to Anthropic, was to submit material favoring one party of the conflict to the UN, presenting it as if it originated from independent local witnesses rather than a state-linked actor.

Targeting European and UN Officials

Beyond the UN, the network compiled detailed dossiers on 18 Members of the European Parliament and several prominent journalists. The company described these as part of a broader strategy to counter accountability mechanisms that have criticized the UAE’s conduct in Sudan. Internal information revealed by Anthropic indicates that the network viewed the concealment of its state affiliation as a key strategic advantage, relying on the appearance of independence to influence political decisions and public opinion in Western institutions.

Uncertainty Over Operational Impact

While Anthropic confirmed it has disabled the specific account used for the operation, the company emphasized the limitations of its investigation. It stated that it could not verify whether the prepared testimonials or dossiers actually reached their intended targets or successfully influenced political outcomes. The report, highlighted by AFP, underscores the growing challenge of distinguishing between genuine civil society advocacy and state-sponsored influence campaigns in the context of the Sudan conflict. The forward question remains whether international bodies will implement new verification protocols to detect such AI-assisted fabrication in future submissions.