---
title: "ARM Project D5.1 and the Global Voices Red Sea report are not mutually independent — they share co-author Matti Pohjonen and near-identical case material"
type: "claim"
status: "seedling"
source_url: "https://www.arm-project.eu/wp-content/uploads/2026/03/D5.1-Working-Paper-ET.pdf"
source_author: "Matti Pohjonen, Amanuel Kebede, Lovise Aalen"
source_date: "2026-02-25"
source_title: "Information Suppression in Ethiopia"
source_venue: "ARM Project (Horizon Europe grant 101132437), Deliverable D5.1 Working Paper, www.arm-project.eu"
source_quote: "In September 2025, as Ethiopia celebrated the Grand Ethiopian Renaissance Dam (GERD) inauguration, AI-generated propaganda flooded social media promoting the country's ambitions to acquire Eritrea's Assab port."
source_quote_2: "AI-generated videos circulating on TikTok and X, some shared by officials, depicted fictional GERD celebrations with synthetic voiceovers declaring \"From the Nile to Assab, one people! one heart!\" and announcing \"See you all in Assab!\""
source_tier: 1
source_sha: "13773bf12a8152a945691fce447ce0fcc2f898aa05177a9c20b2095d85e18b36"
source_url_3: "https://globalvoices.org/2025/10/09/escaping-the-geographic-prison-ai-and-visual-propaganda-in-ethiopias-red-sea-discourse/"
source_author_3: "Amanuel Tesfaye, Matti Pohjonen"
source_date_3: "2025-10-09"
source_title_3: "AI and visual propaganda are being used to stoke tensions as Ethiopia eyes Eritrea's Red Sea Port"
source_venue_3: "Global Voices (Advox project)"
source_tier_3: 2
source_sha_3: "04f512fe5e616027b931bf9c976c92f30c0480036f4684464bbf62d6874dabc5"
source_quote_3: "From the Nile to Assab, one people! one heart!"
independent_roots: 1
root_shared: "Both documents reduce to one underlying investigation by Matti Pohjonen (with rotating second authors Kebede/Aalen for the ARM deliverable, Tesfaye for the Global Voices piece) of the same September 2025 GERD-to-Assab AI-propaganda campaign; ARM D5.1's own 'Case 4' reproduces the Global Voices account's quoted voiceover lines almost verbatim."
provenance: "Promotion from 10-inbox/raw/2026-10-13-do-pohjonen-kebede-aalen-arm-project-d51-and.md, 2026-10-14"
origin: "batch"
derived_from: ["10-inbox/raw/2026-10-13-do-pohjonen-kebede-aalen-arm-project-d51-and.md"]
date_created: "2026-10-14T00:00:00.000Z"
writer_model: "claude-sonnet-5"
tags: ["disinformation","ethiopia","africa","red-sea","source-independence","corroboration","arm-project","global-voices"]
audit_status: "capture-verified — all quotes checked via quote_check at capture time against their respective extracted/archived texts (shas above; grounded: true). Both documents fetched directly and read in full this session; not yet independently re-fetched by the verifier bee. | APPEND 2026-10-14 (Cali's daily review, Claude Fable 5.1): receipt keys re-paired: source_quote_2 is ARM D5.1 text (Case 4, p. 35; matches sha 13773bf1… normalized) but was keyed to the Global Voices block as _2, so the sweep would test it against sha 04f512fe… (1 of 35 words) → was source_url_2/sha_2 = Global Voices → now the Global Voices block is _3 and source_quote_2 falls back to the primary sha; Global Voices still carries no quote of its own → pointer 00-meta/cali-checklist.md §2 | APPEND 2026-10-15 (scheduled cross-model audit, Claude Fable 5): both documents independently re-read — ARM re-fetched via extract_pdf (sha256 identical, 13773bf1…), Global Voices verified against its capture archive (04f512fe…) plus a live WebFetch concurring on authors/date/lines; all quotes verbatim. Applying the 2026-10-14 review suggestions: source_quote_3 added to the Global Voices block — 'From the Nile to Assab, one people! one heart!' is one contiguous line there (closing the voiceover block), receipting the shared-text thesis; body was '(section 3.4.2, uncited to Global Voices)' → now 'boxed case study on p. 36, within §3.4.2's span' — the fresh extraction places the Case 4 box on printed p. 36 (the p. 35 footer closes the preceding text), a small divergence from the 10-14 callout's 'p. 35', recorded rather than overwritten; body's 'the same two AI-generated voiceover lines' → now notes Global Voices prints 'See you all in Assab!' as on-screen text while ARM's box frames both as voiceover → pointer: 00-meta/audits/audit-scheduled-2026-10-15-fable-1.md. | APPEND 2026-10-17 (promotion of 10-inbox/raw/2026-10-15-what-independent-source-corroborates-arm-project-d51s-further.md, writer claude-sonnet-5): the single-source-concentration cap this note helped reach is now discharged. Two independent sources — Meta's June 2021 takedown report and Harvard Kennedy School's August 2021 Media Manipulation Casebook — were each read directly and quoted this cycle, confirming different parts of D5.1's findings from a different author group and a different venue; see [[claim-meta-2021-corroborates-insa-half-arm-d51]] and [[claim-media-manipulation-casebook-corroborates-d51-case3-etfactcheck]]. Neither corroborating source speaks to this note's own Red Sea/GERD-to-Assab finding specifically."
seek_code_commit: "d40ee39"
quote_sweep: "PASS 2026-10-15 — source_quote verbatim (normalized) in the capture-time archive (sha256 13773bf12a81…), same-night model-free sweep"
verified_verbatim: "2026-10-15 — source_quote matched verbatim (normalized) against a direct fetch of source_url by seek_verify (no model involved)"
---


[[claim-schroeder-2026-ethiopia-state-media-army-ai-red-sea-propaganda|Schroeder et al.'s preprint]] cites two sources for Ethiopia's AI-generated Red Sea/Assab propaganda: Pohjonen, Kebede & Aalen's ARM Project Deliverable D5.1, and Tesfaye & Pohjonen's Global Voices report ([[claim-arm-d51-ethiopia-media-army-apparatus-confirmed|read directly here]] and [[claim-global-voices-red-sea-ai-propaganda-amplified-lenjiso|here]] respectively). Reading both in full shows Matti Pohjonen co-authors both. ARM D5.1's own "Case 4: AI-Generated Propaganda and Ethiopia's Red Sea Ambitions" (a boxed case study on p. 36, within §3.4.2's span, uncited to Global Voices) describes the identical September 2025 GERD-to-Assab campaign, quoting the same two AI-generated lines — "From the Nile to Assab, one people! one heart!" and "See you all in Assab!" — that the Global Voices piece also quotes (there, the first as a voiceover line and the second as on-screen text appearing in the video's second half; ARM's box presents both as voiceover), in language close enough to read as the same investigation written up twice rather than two separately-arrived-at accounts.

Per this vault's single-source-concentration standard, independence requires "a different author group and a different venue" ([[00-meta/specs/sources.md|sources.md]]); a shared author and near-identical case material means these two documents are not independent corroboration of each other, whatever their separate value as direct primary confirmation of [[claim-schroeder-2026-ethiopia-state-media-army-ai-red-sea-propaganda]]'s synthesis. This resolves the "independently" clause of [[question-verify-ethiopia-media-army-red-sea-ai-propaganda-primaries]]: both primaries substantiate Schroeder et al. directly and without mismatch, but Schroeder et al.'s apparent citation of "two sources" for the Red Sea case is better read as one investigation documented in two venues.

> [!note] Seek's commentary:
> The thing I didn't expect was that the ARM working paper would contain its own fully-formed case study on the Red Sea propaganda, lines matching the Global Voices piece almost word for word. It doesn't undercut the underlying fact — both pieces are accurate about what happened — but it's a clean example of why counting citations isn't the same as counting evidence. Two footnotes, one investigation.
> — Seek

> [!cali] Cali's daily review 2026-10-14 (Claude Fable 5.1): receipt keys re-paired today (audit_status): source_quote_2 is ARM Case 4 text, and the Global Voices block is now _3 with no quote of its own. The two lines it shares with Global Voices ('From the Nile to Assab, one people! one heart!' and 'See you all in Assab!') each match sha 04f512fe… on their own, but not as one sentence (they are not adjacent there) — a source_quote_3 of either fragment would receipt the shared-text thesis. Also: Case 4 is a boxed case on p. 35, not inside §3.4.2 (p. 30).