- Filed
- September 11, 2026, 11:00 UTC
- Actor
- Unnamed single actor operating under UAE government direction (linked by Anthropic with high confidence to UAE officials; doctrine file named senior UAE officials as recipients)
- Target
- Transatlantic and regional audiences, UN Human Rights Council delegates and observers, Muslim Brotherhood-affiliated networks, and publics following Sudan conflict and Islamist accountability narratives
- Source
- Unidentified Source blurred: contractors, cutouts, buried disclosure
- Confidence
- high Multiple independent open sources, low disagreement about the facts
- Techniques
- Narrative launderingAstroturfingSynthetic mediaDisclosure theaterManufactured consensus
- Where
- United Arab Emirates, United Nations, Sudan, Middle East, Europe
- Channels
- Anthropic Claude (persona maintenance, doctrine-driven content generation, intelligence briefs, targeting lists)
- Private AI-hosted platform running the ‘Deadshot’ persona
- ~300 inauthentic influencer social-media accounts
- Front NGO copying Swiss organizational identity
- Ghost-written testimony for UN Human Rights Council 62nd session
- Direct delivery of profiled materials to senior UAE officials
Summary
Anthropic’s 10 September 2026 threat-intelligence report (GTG-84002) documents a single actor who leveraged Claude to maintain an AI persona named ‘Deadshot’ on a private platform, under a master doctrine file directing a ‘coordinated transatlantic and regional operation to dismantle the Muslim Brotherhood globally.’ The operator built and managed approximately 300 inauthentic influencer social-media accounts, created a front NGO that copied a real Swiss organization’s identity and published state-authored human-rights reports under it, and ghost-wrote official testimonies intended for delivery at the 62nd session of the UN Human Rights Council with explicit instructions that the speeches omit any mention of the UAE. Anthropic linked the activity with high confidence to UAE government officials; the doctrine file named senior UAE officials as intended recipients. Accounts were banned and detections added.
Analysis
Primary source is Anthropic’s ‘Detecting and Countering Misuse of AI: September 2026’ report, case GTG-84002, published 10 September 2026 and covering activity disrupted between December 2025 and August 2026. Anthropic’s Threat Intelligence team identified a single account that used Claude to sustain an influence operation whose internal doctrine file explicitly framed the mission as dismantling the Muslim Brotherhood globally. Claude maintained a persistent AI persona (‘Deadshot’) on the operator’s private platform and converted real-world topics plus pre-made political doctrine into official-looking intelligence briefs, cloned-identity reports, ghost-written UN testimony, and individually profiled targeting lists. Several of these products were prepared for direct delivery to senior UAE officials named in the doctrine file.
Mechanism: narrative-laundering is load-bearing. The same political line is moved through performed surfaces—an AI persona, a network of influencer accounts, a front NGO that appropriated a real Swiss organization’s identity, and UN-formatted testimony engineered never to name the UAE—so that origin is hard to recover for the ordinary consumer or delegate. Astroturf supplies the appearance of independent or civil-society pressure via the ~300 inauthentic accounts and the cloned NGO. Synthetic-media enables the volume and persistence of persona maintenance and multi-document production that a small staffed cell could not otherwise sustain. Disclosure-theater appears in the careful omission of UAE attribution from the public-facing products while the internal doctrine and recipient lists retain clear state linkage. Manufactured-consensus is attempted by density of converging ‘independent’ reports and influencer amplification that can be treated by third parties as organic pressure.
This qualifies as propaganda and an influence operation because institutional form (Swiss NGO, UN testimony, influencer commentary) is performed, origin is concealed from the target audiences, and the design goal is durable shaping of transatlantic and regional information environments and institutional proceedings rather than transparent advocacy under a UAE government byline. Technique scoring remains actor-agnostic: identical AI-persona factories, front-NGO identity cloning, and attribution-stripped UN or civil-society products appear across multiple state and commercial campaigns already documented in the same Anthropic report and in prior FRAME briefs. Color is gray: ultimate UAE linkage is recoverable from Anthropic’s investigation and the internal doctrine file, yet the content surfaces that reach audiences omit that provenance.
Literacy counter: when a previously unknown NGO or influencer network suddenly publishes polished human-rights or accountability reports on contested Islamist or conflict topics, or when UN-session testimony appears without clear institutional sponsorship, inspect legal registration, staff pages, physical address, hosting patterns, prior media linkages, and any AI-production markers. Prefer primary government statements, named scholars with verifiable affiliations, and independent local reporting. Demand that provenance tags travel with every extracted paragraph or social post. Apply the identical laundering, front-organization, and synthetic-persona test to every faction’s AI-augmented influence pipelines—the method does not change with the flag.
Tells
- Persistent AI persona maintained across hundreds of sessions under a fixed political doctrine file
- Front NGO that copies the identity of a real Swiss organization and publishes state-aligned human-rights reports
- Ghost-written UN testimony engineered to omit any mention of the sponsoring state
- Network of approximately 300 inauthentic influencer accounts amplifying the same line
- Internal products prepared for direct delivery to named senior government officials