FIMI & Cognitive Warfare Monitor — 22 November 2026

EU Council sanctions Moscow Social Design Agency (Doppelganger operator), 24 staff and NoName057(16) in largest-ever FIMI sanctions package

Lead Signal

On 16 November 2026 the Council of the European Union sanctioned Social Design Agency, the Moscow-based public relations and political-technology firm that operates the Doppelganger campaign, together with 24 of its staff and six pro-Kremlin commentators, marking the first time the firm has been listed under the European Union own-destabilisation sanctions regime rather than only the Ukraine-war regime. The Council decision names the Russian Presidential Administration as the funder of Social Design Agency and links the firm directly to interference in the 2024 European Parliament elections and to campaigns against Germany and France. In a related but separate designation, the hacktivist collective NoName057(16) was sanctioned under the cyber sanctions regime for distributed denial of service attacks on Ukrainian, United States and European Union government bodies, making this the largest Foreign Information Manipulation and Interference specific sanctions package to date.

The action is significant less for its size than for its legal architecture. By sanctioning a commercial contractor firm rather than only broadcast outlets or individual propagandists, the Council has set a precedent assessed as likely to be reused against other contractor-model operations, most notably Storm-1516. This represents a structural shift in how the European Union categorises Foreign Information Manipulation and Interference infrastructure for enforcement purposes, treating political-technology firms as sanctionable targets in their own right. The milestone lifts attribution and enforcement capacity specifically for the Russia-attributed vector this cycle, even as other components of the information environment, including platform transparency, continue to lag behind.

Other Developments

Matryoshka launches a Spain-specific election interference thread. Ahead of the snap parliamentary election in Spain on 29 November, the Matryoshka network, operating under the banner Operation Overload, launched a campaign assessed as consistent with its established mass-tagging and impersonation tactics, which impersonated El Pais, El Mundo, Agence France-Presse, Euronews, Reuters and France 24 to push Catalan-separatist and European Union collapse narratives. The underlying attribution carries a flagged attribution-over-reach qualification, meaning the campaign is best read as consistent with the documented tactics of the network rather than as a confirmed state-directed operation. The campaign was followed within the same week by the launch of a Spanish fact-checking alliance formed specifically to counter election disinformation through 6 December.

Taiwan refers nearly 2,700 Threads accounts to Meta eight days before its local elections. The Taiwan government flagged the accounts to Meta amid National Security Bureau reporting, again carrying an attribution-over-reach qualification, assessed as consistent with a Chinese information-warfare campaign, with 74 percent of the flagged accounts self-declaring a China location, an unusually direct technical signal relative to typically obfuscated coordinated inauthentic behaviour infrastructure. The referral sits alongside National Security Bureau reporting of a 48 percent year-on-year rise in disinformation volume, exceeding 2.5 million items in 2026, ahead of the 28 November nine-in-one local elections.

France attributes sabotage of approximately 700 retail refrigeration systems to IRGC unit Shahid Kaveh. French Foreign Minister Jean-Noel Barrot attributed the sabotage to the named Iranian Revolutionary Guard Corps unit, with the Paris prosecutor opening a judicial investigation naming individual suspects, a rare formally named high-tier cyber-sabotage attribution carried at Probable confidence.

NATO StratCom quantifies LLM-grooming exposure in AI assistants ahead of the Latvia election. Testing four AI assistants with the same questions in Latvian, Russian and English over ten weeks, the Centre of Excellence found that 34.6 percent of Russian-language answers cited a hostile-state-aligned source, compared with 9.8 percent in English and 4.0 percent in Latvian, while the covert Pravda network itself accounted for only approximately 1 percent of hostile citations, indicating a far broader hostile-aligned source ecosystem than the network alone would account for. A second, independently sourced finding this cycle, the Mediapart and AI Forensics investigation into a campaign designed to manipulate AI chatbot recommendations toward an alternative documentary site, assessed as consistent with links to Israeli state-linked entities rather than directly attributed, reinforces that AI-ecosystem manipulation has moved from qualitative concern toward a measured, reproducible finding category this cycle.

Cross-Monitor Connections

The findings assembled this cycle connect to several sibling monitors. The simultaneous pre-election pressure on the Spain vote of 29 November and the Taiwan vote of 28 November is a direct signal for democratic-integrity, where campaigns assessed as Russia-linked and China-linked respectively are running concurrent electoral interference threads at Probable confidence. The compressed timing of both campaigns around proximate voting windows raises the stakes for coordinated tracking through early December. The three-way contested attribution dispute over the French refrigeration-sabotage incident, running in parallel with heightened Iran, Israel and United States military activity in the Gulf, is a relevant signal for conflict-escalation, tracked at Uncertain confidence pending independent technical verification. The first-ever sanctioning by the Council of Social Design Agency under its own-destabilisation regime, together with the explicit anti-Brussels and European Union-exit narratives documented in the Spain campaign, is directly relevant to european-strategic-autonomy, assessed at Confirmed confidence for the sanctions action itself. Finally, the NATO StratCom LLM-grooming measurement together with the Mediapart and AI Forensics case constitute a first-order signal for ai-governance, documenting independently sourced AI-capability developments within Foreign Information Manipulation and Interference operations at Probable confidence.

Outlook

Over the coming cycle, the Spain and Taiwan elections on 29 and 28 November respectively will close the window during which corrective counter-messaging can meaningfully affect outcomes, making both a primary test of whether platform and government enforcement can keep pace with pre-election campaign tempo. Formal publication of the Viginum confidential report on the Golden Owl French-persona network would allow that attribution to move beyond its current Uncertain rating, closing one of the evidentiary gaps identified this cycle. Publication of a Google Threat Analysis Group bulletin covering the 19,000-channel takedown batch would likewise clarify the actor-specific scope of a sweep that currently rests on a government complaint rather than platform disclosure. Also watch whether the Social Design Agency sanctions template is extended to other contractor-model operations such as Storm-1516 in coming cycles.


Sector & Domain Briefs

On 16 November 2026 the Council of the European Union sanctioned Social Design Agency, the Moscow-based firm behind the Doppelganger operation, together with 24 staff and six commentators, under the own-destabilisation regime for the first time. The Council decision names the Russian Presidential Administration as the funder of the agency and links it to the 2024 European Parliament election environment and to campaigns against Germany and France. This is the largest Foreign Information Manipulation and Interference specific sanctions package to date, and the decision is assessed as a template likely to be reused against other contractor-model operations. The hacktivist group NoName057(16) was separately sanctioned under the cyber sanctions regime for distributed denial of service attacks on Ukrainian, United States and European Union government bodies. Watch for whether this sanctions template is extended to additional Kremlin-linked contractor firms in coming cycles.


Ahead of the snap parliamentary election in Spain on 29 November, the Matryoshka network, operating under the banner Operation Overload, impersonated major European outlets including El Pais, El Mundo, Agence France-Presse, Euronews, Reuters and France 24 to push Catalan-separatist and European Union collapse narratives. The underlying attribution is assessed as consistent with the network documented tactics rather than confirmed as a state-directed operation at this time. A Spanish fact-checking alliance, Comprobado 29N, launched within the same week to counter election disinformation through 6 December. This is the same campaign that has previously surfaced in French riot narratives, and it continues a pattern of narrative persistence and cross-border reuse of established mass-tagging and impersonation infrastructure.


Eight days before the 28 November nine-in-one local elections, the Taiwan government formally flagged nearly 2,700 China-linked Threads accounts to Meta, with 74 percent of the flagged accounts self-declaring a China location, an unusually direct technical signal relative to typically obfuscated coordinated inauthentic behaviour infrastructure. The referral follows National Security Bureau reporting of a 48 percent year-on-year rise in disinformation volume, exceeding 2.5 million items in 2026, with the Bureau documenting AI-generated video and fabricated-poll tactics targeting the island electorate. The underlying attribution to a Chinese information-warfare campaign carries an attribution-over-reach qualification and should be read as consistent with the stated evidentiary basis rather than independently confirmed.


NATO StratCom Centre of Excellence tested four AI assistants with the same questions in Latvian, Russian and English over ten weeks, finding that 34.6 percent of Russian-language answers cited a hostile-state-aligned source, compared with 9.8 percent in English and 4.0 percent in Latvian, while the covert Pravda network itself accounted for only approximately 1 percent of hostile citations. A second, independently sourced finding this cycle, the Mediapart and AI Forensics investigation into the NAZA documentary case, found a campaign designed to manipulate AI chatbot recommendations toward an alternative counter-documentary site, assessed as consistent with links to Israeli state-linked entities rather than directly attributed. Together these findings indicate AI-ecosystem manipulation has moved from qualitative concern toward a measured, reproducible finding category this cycle.

Sources Russia's FIMI and malicious cyber activities: Council imposes sanctions on 30 individuals and two entities → 1 EU sanctions pro-Russia hackers, two dozen staff at Kremlin disinformation contractor → 3 More EU sanctions on Russian propagandists and hackers, as thousands of fake accounts target voters → 3 Peace on Moscow's terms: disinformation ahead of the Abu Dhabi talks → 1 The Russian operation 'Matryoshka' launches a campaign to interfere in the November 29 elections in Spain → 3 Disinfo Update: Member States take on platforms, lessons from Brazil, what's new in Brussels → 2 Taiwan on the frontlines of the fight against Chinese propaganda → 3 Taiwan flags nearly 2,700 China-linked Threads accounts as election influence operation → 3 Indonesia to seek Google clarification over YouTube channel terminations → 3 Taiwan Says Chinese Disinformation Surged 48% Ahead of November Elections → 3 Iran Hacked France's Fridges; Paris Counted Israeli Posts → 3 This Week In Information Warfare 15 - 21 Nov 2026 → 3