Horn Afrik News Agency for Human Rights (HANAHR)

Abu Dhabi Uses Artificial Intelligence to Falsify the Media Narrative on Sudan A new report by Anthropic, developer of the Claude AI system, has exposed a coordinated #UAE-linked influence operation targeting Sudan and UN accountability mechanisms. The findings go far beyond a conventional social-media propaganda campaign: they describe an organized system for manufacturing narratives, impersonating independent institutions, laundering state messaging through supposedly neutral voices, and drafting testimony intended for international bodies. Published on September 10, 2026, Detecting and Countering Misuse of AI examines malicious uses of Claude identified by Anthropic’s threat-intelligence team between December 2025 and August 2026. The Sudan-related material appears on page 46 and, most extensively, on pages 78–80 under case study GTG-84002, explicitly titled: “Disrupting a UAE-directed influence operation targeting the Muslim Brotherhood, Sudan conflict, and UN accountability mechanisms.” According to Anthropic, a single operator used Claude to manage a sustained campaign through an AI persona called “Deadshot.” A central doctrine file repeatedly instructed the system across hundreds of sessions. Claude was therefore not simply used to draft occasional posts; it became part of an operational architecture combining narrative production, political framing, source concealment, target profiling and the preparation of material for political and international audiences. The operation worked through five interconnected channels: Fake-account network: roughly 300 inauthentic influencer accounts were created and managed across social-media platforms. NGO front: the operator cloned the identity of a real Swiss organization and used the front to publish anti-Sudan human-rights material as if it came from an independent institution. Ghostwritten UN testimony: complete statements were authored and drafted for two individuals for the 62nd session of the UN Human Rights Council, with explicit instructions that neither statement mention the UAE. Target dossiers: the operation researched and built detailed profiles on 18 Members of the European Parliament, as well as prominent journalists and other figures, with the apparent aim of shaping or attacking their positions. “Counter-accountability dossiers”: files were compiled on UN Special Rapporteurs who had criticized UAE conduct in Sudan, showing that the campaign targeted not only the public narrative of the war but also individuals and mechanisms involved in international accountability. Anthropic’s most significant finding concerns attribution. Although the campaign was deliberately structured to conceal its origins, the company states that its investigation “linked it with high confidence to UAE government officials.” The central doctrine file also named senior UAE officials as intended recipients of the operation’s products, while the operator funded the social-media network used to amplify them. The report further shows how the campaign sought to manufacture the appearance of independence. The amplification network was centrally funded and coordinated, yet internal reporting described its supposed “independence” as its “greatest strategic asset.” In other words, concealing state direction was itself a core operational objective. That strategy extended directly to Sudanese identities. Anthropic found that the operator borrowed the identity of a real Sudanese human-rights organization and ghostwrote complete UN testimony for two named individuals, allowing material serving a party to the Sudan conflict to reach the UN disguised as the views of independent local witnesses. The draft’s broader political interpretation is that such material reinforced the RSF- and UAE-aligned narrative; Anthropic itself states more narrowly that it served “a party to the Sudan conflict.” Full report: anthropic.com/threat-intelli…