30 OCT 2026 · 25 MIN
Disinformation and Interference in the Grey Zone: Cases and Conclusions

In 2019, I was fortunate to work on the editing of Desinformación. Poder y manipulación en la era digital (Disinformation: Power and Manipulation in the Digital Era), a work coordinated by one of the leading Spanish experts in disinformation, cyber threats, and influence operations, Manuel R. Torres Soriano. Experts such as Guillem Colom, Carlos Echeverría, Nicolás de Pedro, Vicente Vallés, and David Alandete also participated in the book.
That work helped organize and put into perspective phenomena that many of us had already been working on for some time. Seven years later, these phenomena not only remain present, but have acquired new forms, greater speed, and an increasing capacity to transfer their effects from the digital realm to the political, social, and even physical arena.
Revisiting some of the ideas and cases analyzed back then allows us to verify to what extent they remain useful for interpreting recent episodes in France, Italy, the United Kingdom, or Spain. It also helps us understand an underlying logic that remains constant: those acting in the grey zone do not need to create a society’s fractures from scratch; it is enough for them to identify them, exploit them, and thus contribute to turning them into problems of security, cohesion, or governability.
Torres Soriano, Professor of Political Science at Pablo de Olavide University and one of the most recognized Spanish specialists in the study of disinformation and influence operations, has insisted for years on the need not to confuse the two concepts. Disinformation is a tool, whereas influence operations constitute a broader category. In influence operations, true, false, or manipulated information can be combined alongside apparently independent actors, digital platforms, media outlets, proxies, and other forms of intervention aimed at modifying perceptions or behaviors.
The distinction matters because it allows for a better understanding of what is happening. A good part of today’s public conversation continues to reduce the problem to the spread of fake news, as if the threat consisted essentially of a citizen believing false information to be true. The truth, from the perspective of security and the so-called grey zone, is that the issue is much broader. What is relevant is not just whether content is true or false, but what effect it seeks to achieve, what vulnerability it aims to exploit, through what mechanisms it is amplified, and, above all, whether it manages to alter behaviors, erode institutional trust, or turn an existing political or social tension into an increasingly difficult problem to manage.
An Old Phenomenon Adapted to a New Environment
In his analysis of Russian information warfare, Guillem Colom recalled back in 2019, long before the emergence of generative AI, that many of the practices we associate today with bots, social networks, fake profiles, or AI-generated content have antecedents that predate even the existence of the internet. So-called Soviet “active measures” already combined propaganda, disinformation, document forgery, media manipulation, front organizations, agents of influence, and other overt, semi-clandestine, or clandestine tools designed to erode the adversary’s position, polarize their societies, or deteriorate trust in their institutions. That is, there is nothing new in the strategic logic; rather, it is the new technological environment that allows it to be applied today with radically different speed, reach, and costs.
This historical continuity also forces us to avoid a mistaken reading: the grey zone has not just arrived in Europe. European societies have been exposed for many years to operations conducted below the threshold of conventional armed conflict, combining informational, cybernetic, economic, political, and even physical actions that are difficult to attribute (false-flag operations and “little green men,” but also others that are much more sophisticated and hard to identify). What we are observing is a progressive sophistication of these tools and a growing ability to connect them with one another.
In this framework, disinformation gains value not so much because it can create social division out of thin air, but because it makes it possible to identify, reinforce, and sometimes radicalize divisions that already exist. That idea was very present in the book coordinated by Torres Soriano and remains essential for understanding the current logic of influence operations.
A foreign adversary does not need to invent independence movements, immigration, discontent over the cost of living, distrust toward elites, tensions between religious communities, a youth protest, or an institutional crisis. All these phenomena can exist for various reasons and respond to legitimate political issues. Precisely because of this, they constitute particularly attractive surfaces for an influence strategy: the more real the initial conflict is, the easier it becomes to intervene in it without being able to clearly distinguish where spontaneous social conversation ends and where its deliberate exploitation begins.
Therefore, most of the time the objective does not consist of provoking the start of a crisis with a new casus belli, but rather locating an existing crack, widening it, exploiting it, and contributing to making it reach a sufficient dimension to consume public resources, polarize debate, erode trust in institutions, and hinder state action. From this perspective, the success of an operation should not be measured exclusively by the number of people who believed a fake news story, but by its capacity to introduce friction into the political and social system. The qualitative dimension is fundamental in this type of analysis.
Catalonia and the Exploitation of a Pre-Existing Crisis
Spain knows this phenomenon well. The Catalan independence process in 2017 was, among many other things, an example of how disinformation contributed to worsening a crisis, an episode that must be framed within a broader scenario where Russian or pro-Russian actors have tried to exploit internal European conflicts.
The first nuance must be clear: Russia did not create Catalan independence sentiment, and no serious analysis allows explaining the 2017 crisis as the result of a foreign operation. There was a deep-rooted political conflict in Spain that had already reached an extraordinary level of social and institutional mobilization. Precisely for that reason, it offered an exceptional opportunity for actors interested in projecting an image of internal decay in Spain and, by extension, in the European Union.
The analyses carried out at the time detected an intense circulation on social networks of content coming from RT and Sputnik, amplified by anonymous accounts, bots, and other profiles. A subsequent study by the Elcano Royal Institute recalled that during October 1st, information from both outlets featured among the most shared sources under the hashtag #Catalanreferendum, and that a very high proportion of the most active accounts were anonymous and exhibited behavior consistent with automation. Spain, it is worth noting, did not make an official attribution of that activity to a campaign directed by the Russian state.
That caution does not eliminate the relevance of the case. What is important is to observe the mechanism: a real internal conflict could be amplified through narratives presenting Spain on the brink of an irreversible breakdown, exaggerating certain episodes of violence, or trying to transfer the crisis from the plane of Spanish politics to that of a supposed general crisis of legitimacy of European democracies.
From that perspective, the ultimate goal did not necessarily have to be achieving an independent Catalonia, just as it was not necessary for the recipient of those narratives to feel any sympathy toward Russia. A strategic gain could occur simply if a portion of European public opinion ended up perceiving Spain as a state incapable of resolving its conflicts, the European Union as a structure helpless in the face of its members’ internal fractures, or democratic institutions as actors less reliable than they were before the crisis.
Southport: The Information Vacuum as a Vulnerability
The United Kingdom provides one of the best-documented European cases regarding the relationship between disinformation, digital platforms, and physical violence. Following the murder of three young girls in Southport in July 2024, a false identity of the attacker began circulating almost immediately, along with claims that he was a Muslim asylum seeker. The suspect had been born in Cardiff, and the name spread on social media did not correspond to anyone connected to the attack.
Existing legal restrictions on identifying a minor defendant initially limited the information police could provide. The subsequent parliamentary inquiry concluded that this information vacuum created extraordinarily fertile ground for disinformation. Between July 29 and August 9, false or unverified claims about the attacker reached around 155 million impressions on X and were further propelled by algorithmic recommendation mechanisms across various platforms.
The consequence went far beyond a distorted digital conversation. In the days that followed, riots broke out in various cities, including attacks on mosques, establishments linked to asylum seekers, and episodes of violence against immigrants.
The parliamentary inquiry did not conclude that a foreign state was behind the riots. However, it did include testimony from the UK head of counter-terrorism policing, according to whom foreign bots had contributed to extraordinarily accelerating the spread of disinformation. The report also noted that certain patterns detected on the website Channel3Now, which was responsible for initially publishing part of the false information, resembled news laundering methodologies and narrative dissemination used in Russian operations, without establishing a definitive attribution.
Once again, qualitative analysis allows drawing a conclusion from these facts that is more important than any debate over authorship. A real event led to a period of uncertainty; that void was filled by a false narrative (similar to what recently occurred in Burgos); social media algorithms favored its circulation; various actors amplified it; and a portion of that digital mobilization ultimately resulted in physical riots with high levels of violence. That is, this case demonstrates once again that an influence operation does not need to control every phase of the process to reap benefits.
The Particular Appeal of Electoral Processes
Electoral processes constitute particularly favorable ground for this type of operation because, over a relatively short period, they concentrate virtually all variables that favor manipulation: high public attention, intense political competition, uncertainty over the outcome, polarization, strong emotions, and a specific date on which millions of individual decisions carry institutional consequences.
The Fourth Annual Report of the European External Action Service (EEAS) on Foreign Information Manipulation and Interference Threats, published in March 2026, confirms that elections were once again a primary target of Russian activity during the period analyzed. The EEAS documents actions targeting electoral processes in Germany, Poland, Romania, Moldova, and the Czech Republic, among other countries, describing a pattern repeated frequently enough to speak of a true playbook.
That pattern is significant because it is not limited to the timely dissemination of false content. Months before elections, the aim is to establish a presence in the information space, discredit leaders and institutions, and erode trust. As the vote approaches, operations concentrate on divisive issues already present in society, immigration, the economy, energy, refugees, corruption, or protests, and, in the final phase, they may target the legitimacy of the electoral process itself.
However, Russia is not the only actor using these techniques. In France, a particularly illustrative example was observed this year. VIGINUM, the French service dedicated to detecting foreign digital interference, identified an operation named Rokh Solis during the March 2026 municipal elections. The network used coordinated websites and profiles on TikTok, Instagram, X, and Facebook to discredit La France Insoumise and several of its candidates, but also to artificially pit two opposing ideological universes against each other, one identity-based and the other supposedly Islamist, with the goal of polarizing the debate. Investigations found technical indicators pointing toward Israeli actors and, specifically, an influence firm named Blackcore.
The case is especially useful because it reminds us that speaking of foreign interference is not automatically synonymous with speaking of Russia. VIGINUM documented four digital interference operations during those elections and notes that, since the 2017 French presidential election, the country’s major electoral processes have repeatedly been targets of external information manipulation attempts.
Ceuta: When Disinformation Alters Behavior
The assault on the Ceuta border this past summer offers another example of special interest for Spain. A Supreme Court ruling regarding the procedure applicable to individuals entering by sea was progressively simplified and distorted into messages assuring, in essence, that anyone who swam to Spain could not be returned or that the border was open.
The Ministry of Foreign Affairs later had to clarify that the ruling did not modify the Immigration Law, did not prohibit returns, and did not grant any right to remain in Spain or enter the mainland.
Maldita documented that these messages initially circulated with little impact, but on July 30, one of them was shared more than 200 times on Facebook and went viral. In just 24 hours, a massive entry of people into Ceuta occurred (80,000 people, according to the Government of the Autonomous City, whose registered population is 83,595). The same investigation did not find sufficient evidence to claim that a foreign manipulation campaign had caused the mobilization, nor did Spanish intelligence services establish that causal link at the time.
What could be observed was an immediate exploitation of the crisis once it was underway. The pro-Russian network Pravda circulated an old photograph of several Spanish ministers at the Moroccan Embassy to support the narrative that the Government was celebrating while Rabat promoted a new Green March; simultaneously, pro-Israeli profiles pushed messages presenting Ceuta and Melilla as Spanish colonies. The European External Action Service confirmed that Russian attempts to gain an information advantage from the situation had been identified, but also noted that there was no conclusive evidence that the crisis was provoked by foreign state actors.
Ceuta demonstrates with particular clarity why it is useful to separate several levels that are frequently confused. Generating a crisis is one thing; contributing through disinformation to changing individual behavior is another; and exploiting a crisis already underway to introduce narratives favorable to one’s interests is yet another. All three activities can coexist without there necessarily being a single interfering actor or centralized direction over the entire process.
France: From Informational Amplification to Physical Provocation
The events that took place this month in France illustrate the same logic of exploiting a pre-existing fracture in a very different context. The mobilizations of October 17 responded to internal causes within French society and were driven fundamentally by French actors. Indeed, Alliance4Europe’s analysis of over half a million posts related to those protests explicitly states that it found no evidence that a foreign actor was behind their origin. However, it did identify coordinated activity aimed at exploiting existing discontent before, during, and after the demonstrations.
Researchers identified a network they consider very likely linked to Doppelgänger, a pro-Russian influence operation known for creating websites and profiles that impersonate media outlets and for amplifying Kremlin-favorable narratives through coordinated account networks. Over the six weeks prior to October 17, that network spread 639 messages in French, German, English, and Polish, accumulating approximately 7.4 million views. The content drew from real concerns, such as the deterioration of public services or the rising cost of living, to introduce a biased interpretation: that French support for Ukraine was directly responsible for worsening economic conditions at home.
The significance of the episode lies not so much in the figures as in the operational logic. It was not necessary to create a protest against the French government; discontent susceptible to mobilization already existed. What was useful for a foreign actor was inserting itself into that conversation, connecting local grievances with its own geopolitical goals, and contributing to a growing perception that institutions had abandoned their citizens to allocate resources to external interests.
France also offers a more illustrative antecedent of how the information domain can connect with physical actions specifically designed to provoke a social reaction. In September 2025, pig heads were placed in front of mosques and Muslim cultural centers in Paris and its outskirts. Subsequent journalistic investigations linked Serbian perpetrators to a network associated with Russian intelligence services; internal documents obtained by European media showed that the operation was part of an ecosystem close to the Social Design Agency, an organization tied to Kremlin influence campaigns.
The interest in this type of action does not lie primarily in the material damage caused. A pig’s head placed in front of a mosque is, operationally speaking, a simple and inexpensive act. However, it can generate news coverage for several days, provoke political reactions, trigger inter-community tensions, and later feed digital narratives about Islamophobia, insecurity, or the state’s inability to maintain peaceful coexistence. The physical action thus becomes an element of news generation, and the coverage it inevitably attracts forms part of the intended effect.
This is one of the areas where the concept of the grey zone makes complete sense. The boundary between a disinformation campaign, an influence operation, and a physical provocation starts to become much less distinct than those categories suggest when studied separately.
Italy and the Problem of Hasty Attribution
The riots recorded this month in Milan allow us to address the issue from the opposite perspective. Italian Defense Minister Guido Crosetto publicly framed them within what he described as a form of “hybrid warfare” waged against Western societies via TikTok, YouTube, and other digital platforms. However, a day later, Interior Minister Matteo Piantedosi introduced a crucial clarification: authorities had no evidence that the young participants in those riots were being manipulated from abroad.
This caution is not minor. One undesirable consequence of adopting concepts like hybrid warfare, grey zone, or disinformation into political vocabulary is the risk of using them to explain phenomena that may have entirely domestic causes. Not every violent protest is an influence operation, just as a methodology capable of being used by a foreign actor does not in itself prove that such an actor is involved.
This distinction is especially relevant because an exaggerated response can end up producing the very effect it seeks to avoid. If governments interpret real social conflicts exclusively as the result of foreign manipulation, they may neglect the internal causes that originated them, erode the legitimacy of genuine protests, and further damage citizens’ trust in institutions. In short, they may contribute to polarization and the acceleration of the conflict itself.
The problem is precisely that the grey zone is designed to make that differentiation difficult. An external actor can insert itself into a movement it did not create, amplify content produced by real citizens, or use intermediaries who are unaware of the operation’s ultimate goal. Therefore, attribution requires a much higher standard of evidence than simply identifying suspicious behavior.
Barcelona: When the Digital Infrastructure for Mobilization Already Exists
Let us address another even more recent example in Spain: the riots on October 24 in Barcelona, which introduce another relevant variable. The call to assemble under the hashtag Blocus was spread through TikTok profiles, several of them anonymous or recently created, that redirected users to private WhatsApp and Telegram groups. RTVE documented how instructions were shared in some of those spaces regarding clothing, items to bring, meeting points, and even chats discussing theft, looting, and clashes with police.
Following the altercations in Barcelona, which ended with 128 arrests, new calls to assemble began circulating for October 31 in Madrid, Barcelona, Valencia (which resulted in an arrest), Bilbao, Almería, and other cities. EFE Verifica identified dozens of profiles on TikTok, Telegram, and WhatsApp using flyers with similar aesthetics, many generated via AI, presenting highly heterogeneous demands (education, immigration, corruption, or opposition to the Government), yet sharing a common pattern of mobilization and, in some cases, calls to action. EFE notes that indications suggest the phenomenon may not have arisen completely spontaneously, although no promoter has been identified, nor is there currently any attribution linking it to a foreign state.
From a grey-zone perspective, the phenomenon reveals an objective vulnerability: the infrastructure needed to transform a vague narrative into physical mobilization already exists and is extraordinarily cheap. An actor interested in escalating such protests would not necessarily have to create them. It could intervene in a pre-existing structure, amplify certain calls to assemble, introduce more radical content, or artificially boost its visibility. The difficulty in later distinguishing between genuine participants, imitators, provocateurs, and external operators is precisely one of the characteristics that makes these mechanisms attractive. In any case, the net effect increases pressure on the state and its institutions.
Urda and the Information Vulnerability of a Tragedy
The tragic train accident that recently occurred in Urda requires exercising even greater prudence, as we are dealing with an unfolding tragedy with a non-definitive death toll that will in no case be lower than 69, families awaiting information, and an investigation into the causes that has barely begun.
There is no evidence that the accident was the result of sabotage, a cyberattack, or foreign intervention. Despite this, entities like RTVE and EFE have already had to fact-check and debunk posts attributing the crash to alleged Russian or Israeli operations. No authority has raised such hypotheses, and investigations are ongoing.
However, that did not stop social networks and forums from filling up, just hours after the accident, with old photographs presented as current, AI-generated images, videos of other accidents, and conspiracy theories about the causes. Less than 24 hours after the crash, the Data Integration Center had to warn that families of potential victims were receiving fake SMS messages impersonating the Civil Guard to notify them of the alleged identification of their loved ones. This forced authorities to remind the public that such procedures are always conducted in person and never via SMS.
It would be irresponsible at this stage to link those messages to an interference operation. However, the episode demonstrates with harsh clarity how an information vulnerability arises. A tragedy of this scale concentrates uncertainty, pain, an immediate need for information, anger, and extraordinary public attention. During the initial hours, institutions cannot answer every question because many answers simply do not exist yet. That lag between social demand for information and institutional capacity to provide it creates a vacuum that can be filled by errors, rumors, opportunists, criminals, or deliberate manipulation campaigns.
Furthermore, from a grey-zone strategy standpoint, it is not essential to be the author of the initial falsehood. A hostile actor, regardless of its origin and ultimate goal, can observe the ecosystem, identify which rumor is generating the strongest emotional response, and later choose to amplify it or redirect it toward a narrative useful to its own interests. The ability to leverage spontaneously generated content further reduces costs and complicates attribution, as the operation may consist precisely of intervening in a phenomenon whose origin is entirely foreign to the operator.
Disinformation as a Governance Problem
For years, a significant portion of anti-disinformation policies focused on identifying false content and countering it through fact-checking. That task remains essential, but it is insufficient when the objective of an operation is to provoke effects that transcend the informational sphere.
The issue acquires a strategic dimension the moment a narrative succeeds in altering behaviors: when it incentivizes thousands of people to travel to a border believing legal conditions have changed; when it helps turn a tragedy into a violent mobilization; when a provocation against a mosque or synagogue generates days of inter-community tension; when it forces security forces to deploy resources for gatherings of unknown origin and scope; or when it transforms an ordinary electoral process into a debate over the very legitimacy of the system.
At that point, disinformation ceases to be purely a communication issue and becomes a management problem for the state. The intended outcome does not necessarily have to be institutional collapse. It may be sufficient to force institutions to consume resources, political attention, and legitimacy by continuously responding to magnified crises, while public perception grows that the system is progressively losing its ability to control events.
This logic is particularly effective because it operates on issues that are often real. Immigration exists, inequality exists, territorial conflict exists, social discontent exists, and tragedies inevitably generate uncertainty. An effective campaign does not need to convince society of a completely fictitious reality; it can simply limit itself to manipulating the interpretation of true facts, selecting elements that generate the greatest confrontation, and fostering narratives that hinder any consensus-based institutional response.
The Temporal Problem of Attribution
There is also a structural asymmetry that benefits whoever carries out these operations. Democratic institutions need evidence before pointing the finger at a state, an organization, or an intelligence service. A credible attribution can require digital forensic analysis, cross-border cooperation, intelligence work, police investigations, and, on occasion, legal proceedings that drag on for months.
Informational and social effects operate on a completely different timeline. Manipulated content can reach millions of views in a matter of hours, a digital call to assemble can move to the streets in a few days, and a narrative can take root in certain groups before authorities have even managed to determine how it began circulating.
Plausible deniability exploits precisely this disparity in timing. When a specific infrastructure is finally identified, a network of profiles is linked to Doppelgänger, or it is proven that the perpetrators of a physical provocation worked for intermediaries connected to a foreign service, the social effect that motivated the operation may have occurred long before.
This explains why attribution, while essential, cannot be the sole instrument of defense. Democracies must preserve a high standard of evidence precisely to prevent the fight against disinformation from becoming a justification for blaming any internal conflict on external forces. But that should not obscure the fact that, simultaneously, they need to develop prevention and resilience mechanisms capable of acting before knowing with absolute precision who is behind every episode.
The Response Begins with Knowing Your Own Fractures
Here lies probably one of the most relevant conclusions that can be drawn both from Torres Soriano’s 2019 book and from the events observed since then.
Open societies cannot combat the grey zone by ceasing to be open. Nor can they eliminate political disagreements, limit legitimate protest, or turn any institutional criticism into a suspicion of interference. A democracy necessarily encompasses conflict, pluralism, disagreement, and the free flow of ideas. Attempting to eliminate those characteristics to prevent them from being exploited would cause greater harm than the harm it seeks to avoid.
Resilience must be built differently. The first step consists of knowing your own vulnerabilities before actors interested in exploiting them do. That requires identifying which territorial, economic, cultural, or political fractures can become vectors for an influence operation; improving coordination among intelligence, security forces, public administrations, and digital platforms; building capabilities to detect coordinated inauthentic behavior; and developing communication protocols capable of reducing information voids in the early moments of a crisis.
The European dimension is equally important. The operations analyzed show extraordinary ease in moving narratives, methodologies, and digital assets from one country to another. Doppelgänger operates in multiple languages; calls to assemble associated with Blocus use similar aesthetics and techniques across different states; and the election campaigns studied by the EEAS present a common pattern adapted to the specific vulnerabilities of Germany, Poland, Romania, or Moldova. The United Kingdom, although no longer part of the European Union, likewise considers that the threat requires a pan-European response and has strengthened its cooperation with the EU this year against Russian information and interference networks.
It is clear that no anti-disinformation policy can indefinitely substitute for sound management of the issues serving as raw material for these operations. That is, it is necessary to act proactively, not just reactively. If a foreign actor seeks to exploit distrust toward institutions, the best medium-term defense is to reinforce trust through effective institutions. If it tries to instrumentalize a migratory crisis, the state’s ability to provide credible information and properly manage the border reduces the space available for manipulation. If it seeks to question an electoral process, transparent and understandable procedures make it harder for artificially amplified suspicions to find a receptive audience.
That is the least technological and, at the same time, the most important dimension of resilience against the grey zone. Influence operations can be extraordinarily sophisticated, making use of AI, thousands of accounts, cloned websites, and complex distribution infrastructure. However, their effectiveness depends largely on finding a vulnerability within the target society to work on.
The cases of Catalonia, Southport, Ceuta, France, Italy, Barcelona, or the information disorder surrounding a tragedy like Urda are not identical, do not necessarily share the same origin, and, above all, should not be presented as manifestations of a single externally directed campaign. In some, Russian activity has been attributed; in others, foreign intervention from a different source has been established; in others, only subsequent amplification has been documented; and in some, there is no current evidence of interference whatsoever.
Precisely that diversity constitutes one of the main lessons of this phenomenon. The grey zone does not require a single chain of command linking all events. It works best when it can insert itself into dynamics that already exist, leverage spontaneous behavior, use intermediaries, and exploit the difficulty democratic societies face in quickly distinguishing between internal conflict and external intervention.
Therefore, the relevant question is not only whether Europe will be capable of identifying the next Russian, Chinese, or other actor’s campaign (state or non-state, as we have set aside campaigns by para-state entities like ISIS, for example). The deeper question is whether its states will be capable of recognizing their own fractures, managing them, and reducing their potential to become strategic vulnerabilities before someone from the outside figures out how to use them to their advantage. The threat is not new, and the accumulated experience is already extensive. For that very reason, it is time to take action so that the response stops being exclusively reactive.