Center for Strategic Studies of Eurasia

    Commentary · Sep 22, 2026

    Armenia defended its election. Now it needs to show its working

    Russian information operations around Armenia’s June vote were large but mostly aimed abroad. The weak point observers found was the state’s silence about its own counter-disinformation work.

    Security & DefenceSouth Caucasus
    Scenic view of the iconic Cascade Complex in Yerevan, Armenia, under a clear blue sky.
    Mary Rose Relente / Pexels
    Contents

    Key takeaways

    • Civil Contract won 49.75 per cent of the vote and 64 of 105 seats on 7 June 2026, on a turnout of 58.9 per cent, according to the final CEC results.
    • DFRLab found that 91 per cent of the posts in a Russia-linked Storm-1516 campaign against Armenia were in English, and judged the target to be foreign rather than Armenian audiences.
    • In 20 matched Armenian outlets, direct citations of Russian-origin content in mainstream media fell from 44 to 23 per 1,000 articles between the pre-election periods of 2025 and 2026.
    • ODIHR observers recorded that the government rebutted only one false claim on its own Facebook account during the campaign, and that no platform published data on influence operations while it ran.

    Armenia’s 7 June parliamentary election was fought under sustained Russian pressure, and its information space did not break. That is the good news, and it is well supported. The less comfortable finding, in the observers’ own report, is that the state’s defence was conducted largely out of public view by a government whose party was on the ballot. That is the part that now needs fixing, and the EU’s new mission in Yerevan makes it more pressing, not less.

    What happened on the ballot

    Civil Contract, the party of Prime Minister Nikol Pashinyan, took 49.75 per cent of the vote under the final results certified by the Central Electoral Commission on 14 June. Strong Armenia came second with 23.27 per cent and the Armenia Alliance third with 9.92 per cent. Turnout was 58.9 per cent. The allocation gives Civil Contract 64 of the 105 seats. Prosperous Armenia, at 3.99 per cent, fell short of the threshold.

    The international observation mission, led by ODIHR with the OSCE Parliamentary Assembly, PACE and the European Parliament, deployed 457 observers from 42 countries on election day. Its preliminary statement found that voters had a genuine choice in a well-run process. It also recorded an environment in which the Investigative Committee opened more than 129 criminal proceedings after the election was called, most of them involving Strong Armenia supporters, and in which opposition activists in ten provinces and Yerevan described being intimidated, investigated or detained. Both findings belong in any account of the vote.

    The heaviest pressure was not online

    The most direct foreign interference the observers described was economic and diplomatic. On 27 May Russian authorities announced that they would revise the terms on which they supply Armenia with natural gas, oil products and diamonds; the next day they announced temporary restrictions on further agricultural products. On 30 May Moscow recalled its ambassador for consultations, citing steps by the Armenian leadership toward closer relations with the European Union. ODIHR’s head of mission, Janez Lenarčič, described this as unprecedented foreign interference in the form of punitive trade measures. The EU answered on 4 June with a pledge of more than €50 million in immediate assistance.

    This matters for how the information question is framed. A trade embargo announced ten days before a vote is not disinformation. It is a real cost, made visible to voters, with a stated political condition attached. Counter-disinformation tools do nothing about it, and treating every form of pressure as an information problem blurs the line between what fact-checkers can address and what only diplomacy and trade policy can.

    Large operations, mostly aimed elsewhere

    The online operations were real and sizeable. DFRLab, working with CivilNet and Erebusintel, traced 45 coordinated campaigns targeting Armenia to Storm-1516, a Russian state-linked operation, collecting more than 1,650 original posts on X from over 450 accounts. Fifteen of those campaigns drew about 149.5 million views. But 91 per cent of the posts were in English, and the researchers concluded that the aim was less to move Armenian voters than to shape how outsiders saw Armenia and its government. ODIHR reached a compatible finding: it noted an abundance of English-language manipulative content on X, a platform with limited reach in Armenia, and said it did not feature prominently in the domestic campaign.

    A second DFRLab study measured penetration directly. Comparing ten mainstream and ten pro-Russian Armenian outlets over the same months of 2025 and 2026, it found that direct citations of Russian-origin content in mainstream media fell from 44 to 23 per 1,000 articles, while citations of Western sources rose from 60 to about 160. The pro-Russian cluster moved the other way, from 64 to 78. The authors called the Russian presence real but contained. Bots they identified accounted for less than 2 per cent of the comment activity they examined.

    The platforms acted before the campaign. ODIHR records that TikTok removed around 80 Russia-based inauthentic accounts targeting Armenia, Moldova and Georgia by March, and that Meta’s first-quarter report described the takedown of a network it attributed to "Ruposters" that had reached audiences in Armenia among some 20 countries.

    The manipulation voters saw was mostly domestic

    The content that did reach Armenian feeds came, in large part, from the contestants themselves. The observers reported that some candidates and third-party accounts spread AI-generated material to discredit opponents. One recurring opposition message, carried in AI-generated videos, claimed that a Civil Contract win would lead to Azerbaijanis being resettled in Armenia. Senior government figures, for their part, repeatedly accused opposition leaders of allegiance to Russia, and interlocutors raised the use of so-called troll factories by the main contestants. ODIHR concluded that this misleading content reduced the reliability of the information available to voters. None of it required a foreign operator.

    Where the state’s defence fell short

    Armenia did build machinery. A 2025 legislative package on cybersecurity and public information management created a new regulator, though it was not operating by election day. The Information System Agency of Armenia monitored coordinated inauthentic behaviour, the electoral commission had a direct channel to TikTok, and three bodies acted as focal points for reporting to Meta. The Prime Minister’s Office coordinated the government’s response, and an EU Hybrid Rapid Response Team, requested by Foreign Minister Ararat Mirzoyan in February, advised institutions in the spring.

    What was missing was an account of any of it. ODIHR found no public information on the legal and institutional basis for this cooperation or on what the authorities identified and did. The government rebutted exactly one false claim on its own Facebook account during the campaign. No platform published data on influence operations or enforcement while the campaign ran. The observers judged that this limited both the transparency and the effectiveness of the state’s effort.

    The problem is structural, not merely procedural. When counter-disinformation work sits in the office of a head of government who is also leading a party list, and when that government is simultaneously calling opponents Russian agents, opacity lets critics describe the whole effort as campaign activity. The opposition did exactly that: one deputy, Gegham Manukian, called the EU team’s deployment "truly a great disgrace" when it was proposed. Public endorsements of the Prime Minister by some EU and US leaders, which most opposition parties told observers they saw as external involvement on his side, sharpened the point. A defence that cannot be inspected is easy to recast as interference of another kind.

    What to watch

    The EU Partnership Mission in Armenia, established by the Council on 21 April and launched on 13 July, now gives this work a two-year, non-executive home, with a mandate covering cyber threats, foreign information manipulation and illicit financial flows. Its value will depend less on its advice than on whether Armenian institutions start to publish what they flag, to whom and with what result. ODIHR’s final report, which had not been published at the time of writing, will set out formal recommendations; the Venice Commission’s 2024 guidance already points toward an independent, adequately resourced body rather than executive offices. Armenia showed in June that a small state can blunt a large foreign operation. Its next election will test whether it can do so in a way that its own opposition can check.

    Sources

    Related publications

    Reports · Europe

    Georgia's EU Candidacy at Risk as Constitutional Court Upholds Controversial NGO Law

    Georgia's Constitutional Court upheld the Law on Transparency of Foreign Influence on October 17, dismissing challenges filed by fourteen opposition parties and civil society organisations. The ruling, which European institutions and the United States have criticised as incompatible with EU standards, deepens the crisis in Georgia's EU accession process. The European Commission's latest Progress Report, released in parallel, formally recommends the suspension of accession screening for the second consecutive year.