Modern warfare is no longer confined to battlefields, where tanks, missiles and armies fight for control of territory. Alongside traditional conflicts, another front has emerged — one that is less visible but has enormous potential to influence societies and democratic institutions. This battlefield extends across smartphones, social media and digital platforms, where millions of people obtain news, form political opinions and try to understand the events unfolding around them.
The rapid development of artificial intelligence (AI) is fundamentally changing how information is produced, processed and distributed. While the technology offers unprecedented opportunities for communication and access to knowledge, it also provides malicious actors with increasingly sophisticated tools for manipulation, propaganda and interference in democratic processes.
Russia, China and other state and non-state actors have been linked in reports and investigations to information influence operations. Democratic societies are consequently facing an increasingly complex challenge: how to preserve public trust in information when fabricated content can appear entirely authentic.
For decades, propaganda relied on organised structures, institutions, newspapers, radio and television to deliver messages to the public. Today, technology allows an AI-generated video, a manipulated photograph or a statement that was never made to reach millions of users in an extraordinarily short time.
This represents more than an acceleration of traditional propaganda. It changes how a false reality can be constructed — one that may appear just as convincing as a documented event to part of the audience.
A politician can be portrayed in a video saying words they never uttered. A protest that never happened can be illustrated with photographs resembling authentic photojournalism. False information can be distributed simultaneously by hundreds of social media accounts, creating the impression that it has been independently confirmed by multiple sources.
The ability to produce not just messages but apparent evidence is one of the most troubling dimensions of modern information manipulation.
The problem becomes particularly serious during political crises, elections, protests and military conflicts, when citizens demand immediate information and emotions, uncertainty and social polarisation can undermine critical judgement. In these circumstances, fabricated material does not need to remain credible for long to cause harm. Public reactions may occur much faster than facts can be verified.
AI and the expansion of hybrid warfare
This threat must be understood within the broader context of hybrid warfare, which extends beyond military force to include a combination of military and non-military instruments designed to pressure states, weaken institutions and influence political and social decision-making.
Cyberattacks, sabotage, economic pressure, intelligence operations and information manipulation can be combined into a coordinated strategy, with each instrument reinforcing the others to achieve objectives that may be difficult to accomplish through direct military confrontation.
Within this framework, disinformation occupies a distinctive position. Its purpose is not always to change events themselves, but to influence how the public understands and interprets them.
Rather than physically controlling territory, an information operation may seek to shape perceptions, undermine trust in institutions and create an environment in which citizens increasingly struggle to distinguish facts from fabricated claims.
One of the most dangerous aspects of this manipulation is that its objective may not be to persuade people to believe a particular lie. Instead, it may seek to create a persistent atmosphere of doubt in which every piece of information can be challenged and every source portrayed as unreliable.
When authentic photographs can be dismissed as manipulated, genuine videos labelled as AI-generated and professional reporting branded as propaganda, manipulation begins to undermine not just the credibility of individual news stories but society’s ability to establish a shared understanding of reality.
European reports warn of increasingly complex operations
The threat is no longer merely theoretical. Reports from European institutions and technology companies have documented increasingly organised, rapid and complex information manipulation operations.
In its fourth report on foreign information manipulation and interference, published in March 2026, the European External Action Service (EEAS) reported that such operations had reached a new level of complexity and international reach during 2025. The findings reinforced the view that information manipulation cannot be treated simply as a public communications problem.
According to the report, interference in the information environment forms part of a broader hybrid toolkit, in which digital activities can be combined with political pressure, threats and other destabilising actions. In this context, influencing public opinion becomes an instrument for advancing strategic interests.
Russia is among the most extensively documented cases of information influence operations used for geopolitical purposes, particularly since its full-scale invasion of Ukraine. Networks linked to Russian interests have been identified in campaigns seeking to discredit Ukraine, weaken Western support for Kyiv and deepen political divisions within democratic societies.
These efforts go beyond promoting Moscow’s preferred positions. They also include content designed to erode confidence in Western institutions and influence how citizens interpret the war.
The EEAS’s third report, published in March 2025, identified digital infrastructure used by foreign actors, primarily from Russia but also from China, to interfere in the information environments of the European Union and partner countries.
Documented methods included websites impersonating established news organisations, fake online profiles and coordinated distribution of material presented as independent reporting despite being part of organised influence efforts.
Artificial intelligence significantly expands the capabilities available to these operations. It enables actors to produce content in multiple languages, tailor messages to different audiences and distribute material at a speed that previously required substantially more time, personnel and financial resources.
A single narrative can therefore be adapted to different political, cultural and social environments, making its origin and purpose more difficult to identify.
Election interference and the use of deepfakes
An important example of this development emerged during the 2024 U.S. presidential election, when Microsoft’s Threat Analysis Center documented influence attempts by actors linked to Russia, Iran and China.
The activities included fabricated videos, AI-generated content and cyber operations targeting the election information environment. These cases illustrated how emerging technologies were being incorporated into efforts to influence political debate in democratic countries.
However, evidence of an influence attempt should not be confused with proof that it determined an election result. The existence of such operations does not automatically demonstrate that they changed voters’ decisions.
What these cases show is that digital manipulation has become part of organised efforts to influence political debate, shape public perceptions and exploit weaknesses in the information environment.
Another dimension of the threat involves the use of AI to imitate real people’s identities, voices and appearances through deepfake technology. This allows convincing audio and video material to portray someone saying or doing something that never happened.
In a polarised society, where a political statement can trigger immediate reactions, such material can influence public opinion before journalists, institutions or experts have had time to establish whether it is authentic.
Yet the danger extends beyond fabricated evidence. The proliferation of manipulated material creates another troubling possibility: genuine evidence can be dismissed as AI-generated.
Known in disinformation research as the “liar’s dividend,” this phenomenon allows people seeking to avoid accountability to exploit public awareness of deepfakes. As fabricated content becomes more common, it becomes easier to cast doubt on authentic recordings and other genuine evidence.
Information manipulation can therefore operate in two directions simultaneously: making falsehoods appear credible while creating conditions in which the truth itself becomes contestable.
The most serious consequence is not simply the spread of inaccurate information, but the gradual erosion of public confidence in the ability to establish and verify reality.
Automated propaganda and the role of China
AI-powered automation deepens the problem by allowing influence operators to generate thousands of variations of a message, adapting each to the language, culture, concerns and sensitivities of different audiences.
A narrative built around economic insecurity in one country may be reframed around political tensions or distrust of institutions in another, exploiting whichever issues are most likely to provoke a reaction.
In a report published in February 2026 on malicious uses of AI, OpenAI found that threat actors generally did not rely exclusively on AI models. Instead, they combined them with traditional tools, including websites and social media accounts, while frequently operating across multiple platforms and using more than one AI model.
This suggests that emerging technology has not replaced traditional propaganda but has made it faster, more scalable and more adaptable to an environment in which information circulates continuously.
At the same time, producing large volumes of content does not necessarily translate into significant public influence. Security assessments must distinguish between attempts to manipulate, the reach of distributed material and demonstrable effects on citizens’ attitudes or behaviour. This distinction is essential to understanding the threat without exaggerating its power.
China represents another dimension of technology-driven influence. Reports from Microsoft and the EEAS have identified operations linked to Chinese actors that seek to shape international perceptions, exploit social divisions and promote narratives aligned with Beijing’s interests.
Although the methods, priorities and objectives of Chinese operations are not necessarily identical to those associated with Russia, both cases demonstrate that the information environment has become an important arena of competition among major powers.
Why the Western Balkans is particularly vulnerable
These developments are especially relevant to the Western Balkans, a region marked by ethnic tensions, political disputes, external influences and continuing debates over Euro-Atlantic integration.
Historical grievances, identity issues and security concerns continue to shape public debate, creating opportunities for fabricated content to exploit existing sensitivities and generate uncertainty at moments when accurate information is particularly important.
A fabricated statement by a senior official, a manipulated video allegedly showing violence at a protest or false information about a security incident can spread rapidly and provoke reactions before the actual circumstances are established.
In such situations, the problem is not limited to the false content itself. It can also reinforce existing tensions and influence how different communities perceive one another.
Kosovo faces these risks in a particularly sensitive context because of its relations with Serbia, periodic tensions in the north and competing narratives surrounding political and security developments.
Where information about sensitive events can quickly acquire political and ethnic significance, the manipulation of digital content presents a challenge that should not be underestimated.
For example, fabricated material purporting to show an incident between communities, a NATO statement or a decision by Kosovo institutions could generate confusion and uncertainty, particularly if distributed in a coordinated manner and presented as confirmed information.
The consequences could extend beyond ordinary misinformation if false claims influence perceptions of security and shape public reactions to developments that require calm and verified reporting.
Nevertheless, it would be wrong to automatically classify every false story, journalistic error or AI-generated piece of content as part of a hybrid warfare operation.
Identifying such an operation requires evidence of coordination, intent, the actors involved and the way the material is distributed. This distinction is especially important for journalism, which must apply the same verification standards to allegations of information interference as it does to any other matter of public interest.
Journalism must adapt to the AI era
In the age of artificial intelligence, journalists can no longer rely solely on the impression that a photograph looks authentic, a voice sounds convincing or information has appeared on several different websites.
None of these elements, by itself, guarantees that the material is genuine.
Verification requires identifying the original source, checking when and where the material was published, comparing it with independent evidence and examining the context in which it was produced and distributed.
Automated tools designed to detect AI-generated content also have limitations and can produce inaccurate results. They should therefore be combined with other verification methods rather than treated as conclusive proof of a material’s authenticity.
For journalism, this means adopting a more cautious approach to digital sources and developing a deeper understanding of how technology can produce convincing content that does not reflect real events.
At the same time, AI should not be viewed exclusively as a source of risk. Journalists and researchers can also use it to analyse large datasets, identify patterns of coordinated distribution and compare content across different languages.
The same technology that can facilitate information manipulation can also help expose and document it, depending on how it is used and the standards underpinning the analysis.
Defending democracy without undermining free speech
For democratic institutions, the challenge is even more complicated. Efforts to counter organised information interference must not turn the fight against disinformation into a justification for censorship or restrictions on legitimate criticism.
A democratic society needs the capacity to identify and counter operations designed to manipulate public opinion while protecting freedom of expression, media pluralism and citizens’ right to criticise institutions and challenge government decisions.
This is one of the central dilemmas of the digital age. Protecting the information environment cannot be achieved by restricting public debate, just as freedom of expression should not be used as a reason to ignore organised operations intended to deceive and manipulate citizens.
Maintaining this balance requires accountable institutions, professional journalism and a public with access to reliable sources of information.
Ultimately, hybrid warfare in the AI era is not merely a technological contest between states and other actors. It is a broader struggle over trust, authority and the way societies understand reality.
In traditional warfare, strategic objectives may include territory, infrastructure and military capabilities. In information manipulation operations, the target can be citizens’ trust in institutions, the media and one another.
When that trust is continually weakened, the consequences go beyond the circulation of false stories. Even accurate information can lose its ability to guide public debate and help citizens make informed decisions.
A society in which every fact is disputed, every piece of evidence questioned and every information source viewed with suspicion becomes more vulnerable to manipulation precisely because its citizens find it increasingly difficult to form judgements based on verifiable facts.
The greatest danger AI poses to the information environment is therefore not simply that people might believe in an event that never happened or a statement that was never made. It is the possibility of creating an information landscape in which fabricated content and persistent accusations of manipulation lead citizens to doubt even what has been documented and verified.
The battle over information is ultimately being fought not only between those who produce propaganda and those who expose it, but also between efforts to undermine public trust and the need for democratic societies to preserve it.
When citizens lose the ability to distinguish reliable information from manipulation, the damage extends beyond the quality of public debate. It threatens the foundations on which democratic decision-making depends.
In an era when technology can make a lie look like the truth and portray the truth as deception, protecting the credibility of facts is no longer the responsibility of journalism alone. It has become a fundamental issue of democratic security.
