AI Shield trains 60+ Armenian analysts, journalists and civic watchdogs to detect, dissect and answer the AI-powered disinformation operations targeting Armenia's first European parliament. Built by Polish counter-disinfo practitioners. Run by AIISA in Yerevan. Delivered in Armenian.
Armenia stands where Poland stood twenty years ago -with one difference: she has eight months to do it. The Kremlin treats Armenia as 'the next Moldova' and has launched information operations with an unprecedented eight-month lead time. Only 18% of Armenians can identify a fabricated story.
The Matryoshka network kicked off eight months out -a record lead time for Russian influence operations. Over 60 AI-fabricated videos targeting Prime Minister Pashinyan, forged documents, the first-ever audio deepfake of an Armenian PM (June 2025). Storm-1516 -previously deployed against Poland, Germany and France -has expanded its reach into Armenia for the first time.
Source · CSIN monitoring, AIISA report 2026
Since February 2026, the Iran conflict has driven 21.9 million views of forged stories on X, with at least 10 manipulated videos using fake CNN, Reuters and Bloomberg branding targeted at Armenian audiences. 100,000 ethnic Armenians in Iran sit at the immediate edge of this information shock.
Source · Open-source monitoring, March 2026
Azerbaijan demands amendments to Armenia's constitution before signing the peace treaty. It is the perfect manipulable subject: information operations run on a permanent cycle, with narratives that target the government, the opposition and the European turn simultaneously.
Source · CSIN/AIISA analysis
Fact-checking verifies single claims. Operations analysis identifies coordinated campaigns. Without that capacity, Armenian CSOs see the trees but not the forest.
AI Shield builds capacity on two levels at once. The decisive innovation: Track A graduates become Track B trainers. If AIISA's trainers can train the next cohort, the transfer worked.
One trajectory: diagnose, train trainers, cascade, document. Every step is measurable and dated.
10-12 August 2026 · 3 days · Yerevan · PLN 13,960
Three days of high-intensity diagnostics with 15 analysts from AIISA and invited think tanks. Poland's road from target to exporter of resilience (2014–2026). Narrative parallels between the Armenian 'Ukrainisation' frame and the 'NATO means war' frame deployed against Poland. Mapping the actors: Russia (Matryoshka, Storm-1516), Azerbaijan, Iran (IRGC). Survey of AI tools and curriculum design for Activity 2.
Outputs · Threat-landscape map · Curriculum · Tool inventory
October–December 2026 · 54 hours · Yerevan + online · PLN 22,140
In-person block in Yerevan, 5 days × 6 hours: AI-powered OSINT (Junkipedia, Media Cloud); narrative analysis and coordination detection on Storm-1516 (Botometer, Bot Sentinel); deepfake detection (InVID WeVerify, FotoForensics, Hive, TrueMedia.org); social network analysis (Gephi, NodeXL); practical integration and trainer preparation. Online block, 6 weeks × 4 hours: each participant picks a real campaign targeting Armenia, analyses it and writes a report, mentored by CSIN experts.
Outputs · 12–15 reports · Certified analysts · Trainers ready
January–March 2027 · Yerevan · PLN 14,920
Three 16-hour cycles (10–12 participants each) for staff from CSOs and independent media. AIISA trainers, certified in Activity 2, lead the workshops; CSIN only provides online mentoring. This is the empirical evidence of transfer -not a declaration, but a measurable demonstration that the partner can train the next cohort on its own.
Outputs · 30–36 trained · Implementation plans · +25% measured
April–May 2027 · Online + Yerevan (hybrid) · PLN 15,960
Open-source 'AI Shield' Toolkit -~60 pages, three languages (PL/EN/AM), CC BY-NC 4.0. Five chapters: typologies of coordinated information operations, Poland's road from target to exporter, step-by-step tooling, Armenian case studies, action plan for newsrooms and CSOs. Analytical report distributed to Armenia's MFA, parliamentary committees, the EU Delegation. Expert panel in mid-May 2027 -hybrid format, minimum 50 attendees.
Outputs · Toolkit PL/EN/AM · Analytical report · Launch panel
No cost barrier to continued use after the project closes. Each tool documented in the open-source Toolkit, in three languages.
JunkipediaMedia CloudCase study
Mapping the Matryoshka network -from a single AI-generated video to the full campaign infrastructure.
BotometerBot SentinelCase study
Analysing Storm-1516: account-synchronisation patterns, narratives, channels.
InVID WeVerifyFotoForensicsHive ModerationTrueMedia.orgCase study
The first audio deepfake of PM Pashinyan -the full verification trail.
GephiNodeXLCase study
Visualising the coordinated-account network in the 'Ukrainisation' campaign against Armenia.
The AI Shield team brings together Polish practitioners of information-trust reconstruction and Armenia's security research community. We don't write about disinformation -we analyse it, and train the people who answer it on the ground.
CSIN · Caucasian Social Innovation Network · Gliwice
President, project coordinator
Analyst at INE PAN and ISP PAN (Three Seas Initiative Research Centre). Polish MFA diplomatic posting in Tbilisi. Former Frontex. Blavatnik School of Government (Oxford), PISM and European Academy of Diplomacy fellow. Law graduate, University of Silesia.
Project roleCoordination, leading Activities 1–2, co-author of the Toolkit.
AI expert
Data analysis and applied-AI specialist for civic organisations. Of Armenian heritage -full command of Armenian cultural, political and linguistic context.
Project roleCo-delivery of AI modules in Activities 1–2, lead author of the Toolkit's tools chapter, Armenian-language mentoring for AIISA trainers in Activity 3.
Training coordinator, President of PolandAI
Digital-competence trainer with nearly 1,000 instructional hours and 1,500+ people trained. KPO programme lead (6,386 local-government officials), trainer for Obywatel.IT (Polish PM's Office and Ministry of Digital Affairs), facilitator on dozens of Erasmus+ projects.
Project roleInitiated contact with AIISA in autumn 2025. Co-delivery of Activities 1–2, logistics coordination, mentoring support for AIISA trainers in Activity 3.
AIISA · Yerevan
Director, partner coordinator
Former Member of Armenia's National Assembly (2007–2012), delegate to the NATO Parliamentary Assembly. Former chair of Armenia's Public Council (2019–2021). Senior expert at Agora Strategy Group (Berlin). Munich Security Conference, Chatham House and Wilton Park participant. Co-lead of the ongoing NATO SPS project 'Guardian AI' (EUR 80,000).
Project rolePartner coordination, security expertise, recruitment, organisation of Activity 3 and the expert panel.
Permanent research team · Track A core
Track record: Visegrad Fund 2023, NATO SPS 'Guardian AI' (ongoing), a decade of NED partnership.
Project roleBackbone of Track A. Peer review of participant reports. Local recruitment and logistics.
How AI Shield came to be
In autumn 2025, Adrian Misiak -while researching the NATO Science for Peace and Security programme -identified AIISA as the leading Armenian analytical centre combining security expertise with hybrid-threats experience. A series of working meetings with Director Safaryan from October to December 2025 confirmed the complementarity: CSIN brings AI methodology and training, AIISA brings security expertise and the Armenian network. AI Shield is the first joint project. CSIN and AIISA plan a joint NATO SPS application in 2027.
Three layers of sustainability. Every result has a pre/post measurement, a date and a public artefact.
Twelve to fifteen certified analysts publish their own analytical reports.
Pre/post test ≥ +30%.
Thirty to thirty-six CSO and media staff walk out with individual implementation plans.
Competency gain ≥ +25%.
Three workshop cycles delivered autonomously by AIISA -without CSIN in the room.
Participant rating ≥ 4/5.
An open-source Toolkit in Polish, English and Armenian -CC BY-NC 4.0.
Two hundred-plus downloads, twenty-plus recipient organisations.
An analytical report for Armenia's Ministry of Foreign Affairs, parliament and the EU Delegation.
At least two citations in media or institutional reports.
Institutional
AIISA carries the capacity built in AI Shield directly into the NATO SPS 'Guardian AI' project (EUR 80,000).
Individual
Certified analysts use AI tools in their daily research and editorial work -beyond the project's funding cycle.
Systemic
Armenia will go through more electoral cycles, a constitutional debate and a possible accession referendum. Each generates a wave of disinformation. AI Shield lays the resilience foundation without which European integration cannot be defended.
NATO SPS 2027.
Joint CSIN–AIISA application to the NATO Science for Peace and Security programme. AI Shield builds the institutional trust foundation that makes the bigger application credible.
Regional replication.
The English-language Toolkit enables adaptation in Georgia, Moldova and Ukraine -with no translation cost. Russian information playbooks travel; defensive playbooks should travel faster.
'AI Shield Georgia'.
Project experience provides the basis for a subsequent RITA or Erasmus+ KA220 application focused on Georgia, applying the AIISA-style cascade model in a second post-Soviet democracy.
"The success or failure of Armenia's European integration depends in part on whether its civil society can defend itself against information operations designed to reverse that course."
AI Shield · AIISA · 2026-2027