It appears difficult to dispute that the online public spheres in the EU’s member states are today confronted with a constant stream of deliberate political manipulation and deception campaigns. This is attested by a long stream of recent high-profile controversies, from the sudden appearance of various imitation news websites containing false reporting after Russia’s invasion of Ukraine inter alia in Germany and France (‘Operation Doppelgänger’), to the Romanian Constitutional Court’s decision to annul a round in the presidential elections due to Russian interference aimed at promoting the outsider hard-right candidate Georgescu, and from much-publicised but orchestrated anti-Semitic displays in Paris’ streets, to a baseless Islamophobic conspiracy outbreak after an Amsterdam church burned down last New Year’s Eve. Such campaigns do not remain without effect, frequently leaving those exposed to them epistemically confused and disoriented, discouraged and withdrawn from public debate, or polarised and radicalised on specific ‘trigger’ issues.

Over the past decade-and-counting, evidence of the structural involvement of third-state actors in these artificially fomented controversies has been accumulating. When in 2014 Finnish journalist Jessika Aro began to systematically investigate the activities of St. Petersburg-based ‘troll factories’, she quickly faced intense harassment and character assassination from pro-Kremlin propagandists. An early and prominent organisation among these ‘troll factories’, the so-called Internet Research Agency, has been linked to the Russian government. In 2024 Operation Doppelgänger was unmasked and countered by various governments including those of Germany and the United States as a Russian-government commissioned covert operation. And most recently, online forensics research consortia laid bare systematic interference in Dutch and Hungarian national election debates by the same troll farms based inter alia in Nigeria, suggesting Russian involvement to be ‘probable’.

Understandably, the European public is beginning to express increasing concern. 48% of respondents to a recent Dutch survey stated being ‘reasonably to very worried’ about misleading information on the internet. Similar worries exist across Europe, appearing particularly pronounced in Southern Europe. Not only ordinary citizens express deep concern. Experts, too, place ‘misinformation and disinformation’ very high in a ranking of most severe global risks, both in the short term (second place) and long term (fourth place).

In the face of a predicament (both real and perceived) of this magnitude, it appears only reasonable that the European Union would seek to come up with decisive measures. Over the past decade, it has indeed built up a policy capacity and toolbox for doing so. A primary focus in this regard is on Foreign Information Manipulation and Interference (FIMI), a concept developed by the European External Action Service (EEAS) starting from around 2023. The EEAS describes FIMI as “a mostly non-illegal pattern of behaviour that threatens or has the potential to negatively impact values, procedures and political processes”, leading to “activity [that] is manipulative in character, conducted in an intentional and coordinated manner, by state or non-state actors, including their proxies inside and outside of their own territory”.

Several aspects stand out in this definition. First, the identification of FIMI requires a truth assessment, establishing to a non-trivial standard that circulated information was indeed manipulated. Second, to meet the standard of FIMI, such manipulation must be underpinned by a real harm motive, in which there is an intention to undermine European “values, procedures and political processes”. Third, a sufficient causal ascription is required, to the extent that driving actors with demonstrable intentions are identified. Each of these three criteria are forensically challenging to prove and vulnerable to epistemic debate in the most straight-forward of disinformation cases, but exceedingly complicated where foreign actors and their proxies collude in sophisticated covert operations. Such instances would after all be characterized by plausible deniability and attempts to stay below the regulatory/criminal legal threshold. Indeed, one might wonder what the regulatory purpose is of a taxonomy of behaviours that are “non-illegal” as the EEAS’ own definition contends, in a context that is characterized by robust fundamental rights guarantees with regard to the freedom of opinion and speech under Article 11 EUCFR and the Article 10 ECHR, the scope of the latter expressly extending “regardless of frontiers”.

Aside from definitional and rights-based limits, the literature on counter-disinformation strategies suggests that the development of effective measures against the harms of disinformation is exceedingly challenging. Tailored to situations, audiences, and social media platforms, disinformation takes on many forms, from brazen agitprop to subtle psy-ops. Typically, the intention and identity of disinformation originators remain elusive, and ideologically steeled ‘foot soldiers’ or unwitting, ‘misinformed’ individuals play a key role in spreading their message. What’s more, in response to online trends and algorithm adjustments, disinformation strategies constantly evolve.

In the face of these overwhelming challenges, each of the most common approaches to minimising the harms of disinformation (fact-checking, financial investments, media literacy training, and due diligence obligations) has its shortcomings. Thorough fact-checking and debunking is often a time-intensive affair, meaning that the ‘straightened record’ often trails the original false message to the extent that much of the damage of manipulation of public perception and opinion will already have been done when the fact-check appears. Worse still are situations in which fact-checking turns counter-productive by leading false messages to reach previously untapped audiences, or becoming itself subject of controversy. This was the case when in 2018, a legitimate Dutch regional newspaper accidentally ended up on the blacklist of disinformation cases of ‘EU v Disinfo’, a counter-disinformation initiative sponsored by the EEAS sparking a forced retraction and significant reputational damage to this EU-sponsored instrument. Financial investments in the media ecosystem, while likely having a positive effect on quality and epistemic pluralism, are typically very costly and subject to long effect horizons. Moreover, they can be perceived as arbitrary, as was the case with a €120 million low-interest loan granted in 2021 to DPG Media, a dominant media group of legacy newspapers. In the case of media literacy training, the academic debate remains out about the size of the ‘treatment effect’ of those trained. In any event, ‘prebunking’ methods such as the EU-subsidised serious game ‘Bad News’ lack a lasting effect, thus requiring regular repeat trainings. A more hands-on tool such as due diligence obligations at pain of a fines system must be carefully calibrated to avoid causing a chilling effect on free speech. This would seem a particular risk where such obligations are content-oriented, as is the case with the German Network Enforcement Act.

Yet the most invasive form of regulation is undoubtedly that of (administrative or criminal) repression of information or senders. Disinformation-related restrictive measures come under this category. Not only do such measures quickly risk becoming sweeping and disproportionate, they also have questionable effectiveness. After all, in a social media landscape characterized by low-cost and multi-channel dissemination, (false) information is highly portable and can thus be quickly rerouted when specific outlets are banned (see Lonardo’s comments on the case C-67/25, Traugott Ickeroth, in his post here). Particularly where the circulation of banned but otherwise non-illegal content reverts to (closed) social media channels, this can lead to a ‘cat and mouse’ game in which effective removal from the public sphere would require a level of repression that far outweighs its realized benefits. After all, much non-illegal but leading/suggestive communication seeks out pre-existing societal conflict for wider uptake. Particularly when lacking meticulous justificatory substantiation, banning such communication may have secondary trust-undermining ‘Streisand’ and ‘conspiratorial’ effects, that serve as a badge of honour to their senders and add further confirmation to their delegitimate subtext.

Posted by Maarten Hillebrandt (Utrecht University)

---

This post is the first one in a series entitled ‘The EU’s restrictive measures against Russian media outlets and ‘foreign information manipulation and interference’: Issues of effectiveness, legality and principle’. The Introduction can be found by clicking on this link.