Czech MEP links AI fakes to anti-Polish disinformation campaign

Disinformation campaigns targeting Central European bilateral relations frequently exploit unresolved or sensitive historical territorial adjustments to test alliance resilience. The Cieszyn Silesia region has historically served as a flashpoint for nationalist sentiment due to border shifts following World War II and subsequent treaties. Influence operations in this domain typically follow a pattern of 'historical grievance amplification,' where dormant disputes are reactivated through digital fabrication to create the appearance of contemporary diplomatic crises. Prior incidents involving NATO member states have demonstrated a recurring tactic of mimicking official government communications and public broadcasters to lend false legitimacy to irredentist narratives. These operations often coincide with periods of heightened regional security cooperation, aiming to introduce friction into logistical and political support networks. The integration of generative AI into these workflows represents an evolution in tradecraft, allowing for the rapid production of localized visual and textual content that targets specific linguistic and cultural demographics. Historically, attribution in such cases has lagged significantly behind detection, creating a window where political responses must rely on probabilistic assessments rather than definitive forensic proof.

On August 25, 2026, Czech Member of the European Parliament Tomáš Zdechovský publicly alleged the existence of a coordinated disinformation campaign utilizing artificial intelligence to manufacture friction between the Czech Republic and Poland. According to posts made on X (formerly Twitter), Zdechovský identified specific synthetic media artifacts, including AI-generated content and fake social media profiles, as central components of an operation designed to fabricate events such as a protest in the city of Třinec [1][2]. The MEP asserted that the strategic objective of this campaign is to alienate two NATO allies from one another, explicitly stating that "the trail leads East," though he did not provide forensic attribution or name a specific state actor in these initial warnings [1][2]. Zdechovský tagged POLITICO Europe in his alert, effectively using the outlet as a dissemination channel for the allegation, although no independent verification of the claims was presented in the source material at the time of posting [2].

Security analyst Lukasz Olejnik provided a detailed technical breakdown of the alleged operation that both corroborated and nuanced Zdechovský’s claims. According to Olejnik, the campaign exploits a genuine historical grievance regarding a post-1958 territorial adjustment involving approximately 368 hectares in the Cieszyn/Těšín Silesia region [3]. Olejnik stated that the disinformation operation falsely reframes this settled "territorial debt" as an imminent Polish land grab to inflame nationalist sentiments. The methodology described by Olejnik includes the creation of fake diplomat accounts, the production of AI-generated material, the forgery of Polish government documents, and the cloning of the Czech public broadcaster iROZHLAS website [3]. Additionally, Olejnik noted that a Radio PiK page was hacked and used to promote a fictitious protest in Třinec via paid advertising [3].

While Zdechovský characterized the threat as an active attempt to turn nations against each other with Eastern origins, Olejnik offered a more measured assessment of the campaign's immediate impact and attribution. According to Olejnik, the reach of the operation appears limited, but the primary risk lies in the potential for nationalist mobilization and the erosion of trust between allies supporting Ukraine [3]. Crucially, while acknowledging that Russian involvement is "plausible suspected," Olejnik clarified that no formal attribution had been made linking the campaign to Moscow or any other specific entity [3]. This distinction highlights a gap between the political alarm raised by Zdechovský and the evidentiary standards typically required for definitive geopolitical attribution.

The sequence of events suggests a rapid response to emerging information operations targeting Central European cohesion. Zdechovský’s posts serve as a political signal intended to preemptively inoculate the public against synthetic narratives, leveraging his platform to label the content as manipulative before it achieves organic virality. By citing specific markers like the Třinec protest and AI generation, the MEP attempted to provide tangible identifiers for the public and media to recognize the alleged fakes. However, the reliance on social media platforms for both the dissemination of the disinformation and the subsequent debunking creates a complex information environment where allegations of AI-driven interference must compete with the very content they seek to counter. The involvement of cloned legacy media sites like iROZHLAS indicates a sophistication in the alleged campaign that goes beyond simple bot networks, suggesting an attempt to borrow institutional credibility through digital mimicry [3].

As of the available timeline, the controversy remains centered on the identification and characterization of the campaign rather than on confirmed consequences or adjudicated responsibility. The allegations present a case study in how political figures and security analysts navigate the intersection of historical territorial disputes and modern generative AI capabilities. The narrative established by Zdechovský frames the incident as a direct attack on alliance unity, while Olejnik’s analysis grounds the event in specific informational tactics and historical context. Both parties agree on the presence of synthetic elements and the anti-Polish/Czech divisiveness of the content, yet the precise origin and ultimate efficacy of the operation remain subjects of assessment rather than settled fact. The situation underscores the challenges of verifying AI-driven influence operations in real-time, where the speed of political warning often outpaces the tempo of forensic confirmation.

The primary stakeholders are Czech and Polish policymakers and security services responsible for maintaining bilateral trust within NATO structures. The risk involves the degradation of political solidarity and public support for mutual defense commitments through manufactured crises. While the immediate audience reach appears limited according to analysts, the exploitation of a specific 368-hectare territorial dispute demonstrates a precision targeting of historical trauma that could catalyze nationalist fringe movements. The magnitude is currently defined by reputational and diplomatic friction rather than kinetic or financial loss, but the precedent of using AI to clone public broadcasters like iROZHLAS raises the baseline cost of future verification efforts for all regional media consumers.