Saturday, August 29, 2026

News

Icelandic Police Find No Evidence of Russian Interference in EU Referendum

ResearchPatryk Raba
Icelandic Police Find No Evidence of Russian Interference in EU Referendum
Fot. Bragi Þór Jósefsson / Althing, Wikimedia Commons (Attribution (licencja Althingi))

Despite months of warnings from Iceland's government about Russia, police say they have found no evidence of foreign interference in today's referendum on resuming EU talks. Researchers and disinformation watchdogs say fake content is being produced domestically, with cheap AI-generated bots emerging as a new threat.

Contents
  1. Ministry Warnings, Police Findings
  2. Trail Leads Home, Not Moscow
  3. Cheap AI-Built Bots
  4. Oligarchs as an Alternative Threat

On Saturday, August 29, Icelanders are voting in a referendum that will decide whether the country resumes EU accession talks, suspended since 2013. Throughout the summer, the government in Reykjavik warned that the campaign could be disrupted by Russian disinformation and internet bots. Police, however, admit they have found no evidence of this, while researchers tracking the Icelandic internet point instead to domestic sources of fake content and cheap, AI-generated bot accounts.

The referendum is exceptionally close. An August poll by the Maskína institute put support for resuming EU talks at 53 percent versus 47 percent opposed, but the latest Gallup poll, conducted between August 24 and 27, showed the anti-integration camp ahead for the first time in the entire campaign. Prime Minister Kristrún Frostadóttir has described the vote as a now-or-never moment. The issue also divides the island's Polish community of more than twenty thousand, the largest immigrant group in a country of fewer than half a million people.

Ministry Warnings, Police Findings

Foreign Minister Þorgerður Katrín Gunnarsdóttir has for months spoken of foreign efforts targeting the referendum debate, comparing the risk to Britain's Brexit. As an example, she pointed to a wave of fake accounts that mass-commented on posts by businessman Alfreð Tulínius, criticizing him while praising the minister herself, with comments written in unnaturally machine-like Icelandic.

Icelandic police have nonetheless been monitoring the situation since July, and, as Deputy Commissioner Finnbogi Jónasson told the outlet Vísir, criminal intelligence is tracking suspicious social media accounts, sites of unclear origin, and possible foreign funding of the campaign. Even so, by mid-August investigators had disclosed no evidence confirming interference by any specific foreign state in the vote.

Trail Leads Home, Not Moscow

Icelandic disinformation researchers have long argued that the biggest problem isn't foreign troll farms but content produced locally. Eiríkur Bergmann, a political scientist at Bifröst University, notes that around 40 percent of Icelanders say in surveys they believe in the existence of a deep state, even though the concept itself was imported from foreign discourse. In his view, local disinformation nonetheless follows its own, domestic trajectory.

The Icelandic disinformation we've seen so far has been homegrown - Eiríkur Bergmann, political scientist, Bifröst University

Cheap AI-Built Bots

Netvís, a media literacy organization, documented automated accounts as far back as the last local elections and warns that the cost of deploying similar tools has fallen sharply thanks to generative AI. Icelandic, a language spoken daily by fewer than 400,000 people, used to be a natural barrier against mass disinformation from abroad. But new language models can now write fluent Icelandic, removing that barrier. Netvís has launched an educational portal explaining online manipulation methods, including deepfakes, bots, and emotionally targeted content, and works with police and social platforms to report the most serious cases.

We have every reason to be aware that this technology can be used to stoke conflict and influence our views - Stefanía Ragnarsdóttir, Netvís

Oligarchs as an Alternative Threat

Retired University of Iceland economics professor Thorvaldur Gylfason told EUobserver that the real risk of the referendum being hijacked lies elsewhere, not in Moscow. He said the anti-integration camp has a considerably larger budget than supporters of EU talks and has brought in foreign experts, including Swiss lawyer Carl Baudenbacher, previously known for work tied to circles close to the Russian government. It's a paradox of the campaign: a figure linked to Russia is working not for the Kremlin but for the domestic anti-EU camp.

For Polish readers, the Icelandic case is a warning that arrives ahead of their own experience. A country with a small population and a language used almost exclusively at home was long considered resistant to mass disinformation from abroad precisely because of that language barrier. AI models capable of writing fluently in rare languages are removing that barrier, shifting the weight of the threat away from foreign intelligence services and toward domestic political and business actors, who can now use the same tools more cheaply than ever.

The referendum result is expected on Saturday night into Sunday, Icelandic time. Regardless of the outcome, Icelandic police and Netvís say they will continue tracking accounts and content tied to the campaign, and the question of where disinformation originates in an age of cheap AI tools is expected to resurface at every future election on the island.

Share: