Cognitive warfare in action: how a fabricated Poland–Czechia border row reached real news pages
Last week was a pretty interesting one: unknown operators staged a Poland–Czechia border row and planted it on a hacked public radio site, a Russian bot network buried Sweden’s fact-checkers in fake videos ahead of the election, and Russians running a fake Israeli think tank gave themselves away through their own ChatGPT prompts. Welcome to issue 14 of Narrative Ops.
A forged letter started a border row between Poland and the Czech Republic
Over several weeks, a coordinated operation staged a territorial dispute between Poland and the Czech Republic. It began on Facebook, where an account posing as the private profile of Czech Ambassador to Poland Břetislav Dančák published a forged Polish foreign ministry note demanding the return of Czech land, CyberDefence24 reported. The same account posted an AI-generated video in which “Dančák” condemned the supposed Polish demands, and other accounts impersonated Polish Deputy Foreign Minister Artur Harazim.
The operation then moved into real media: hackers stole a journalist’s password, broke into the website of Radio PiK, a regional public broadcaster in the northern Polish city of Bydgoszcz, and published an article claiming Poland was preparing to occupy over 600 hectares of Czech territory. The station deleted the article and reported the breach to Poland’s Central Bureau for Combating Cybercrime.
The attackers then built a fake Czech newsroom: a copy of the public broadcaster’s news site iRozhlas.cz, running on lRozhlas.cz (one letter different). Its article claimed Czechs in the affected areas were already selling property out of fear. Czech Radio had the site taken down – the domain was registered through a German registrar and hosted in Turkey – and says it is considering further legal steps.
In the Czech Republic the operation moved to Facebook. A profile called Adam Sikora, set up a fortnight earlier with a portrait that a detection tool flagged as AI-generated, paid to promote a demonstration on Třinec’s Freedom Square. Sikora also advertised a house for sale in nearby Ropice: Facebook labelled those photos as AI-generated, and Halina Szczotka, editor of the local monthly Zwrot, told Seznam Zprávy no such house exists there.
There is a real dispute underneath: a 1958 agreement adjusting the border left about 368 hectares in Czechoslovakia’s favour, which Warsaw treats as an unresolved territorial debt and pursues through diplomacy. The fabricated articles nearly doubled that figure, putting the supposed Polish demand at 638.44 hectares. NASK, Poland’s state cybersecurity institute, is documenting the campaign and told CyberDefence24 the methods were “consistent with methods known to be used by the Russian side”, while making no formal attribution. In the Czech Republic the foreign ministry is examining the case, and opposition MP Andrea Hoffmannová has filed a criminal complaint together with Pirate Party analyst Janusz Konieczny.
A Russian bot network is targeting Sweden’s election by burying the fact-checkers
Seven fabricated videos about the Swedish elections appeared on X within one hour on 18 August, marked with the logos of national media including SVT and Dagens Nyheter, as well as other European media - Euronews and Der Spiegel. Prime Minister Ulf Kristersson linked the activity to Matryoshka, a Kremlin-connected network that makes fake content dressed up as reports from well-known news brands. “We see this across Europe. The Russians are targeting countries that support Ukraine,” Kristersson told SVT.
According to analysts, the campaign was not primarily designed to reach large audience on social media, the aim was different. Operators send the fake posts to journalists and fact-checkers, usually disguised as requests to check whether something is true. According to disinformation experts quoted by SVT, this may be meant to tie up newsroom verification capacity. The campaign may also aim to produce coverage that spreads the messages further or makes the campaign appear more influential than it is. Most of the posts concerned Western military and financial support for Ukraine.
Sweden’s Psychological Defence Authority issued a separate warning on 25 August. Jerker Sundstrand, who runs the agency’s election work, said Russia is running information operations against Sweden through public statements and through networks spreading disinformation, and that the risk rises in election years, Norran reported. The agency pointed to similar patterns around votes in Moldova, Germany and France.
Russians built a fake Israeli think tank, and their ChatGPT prompts gave them away
On 25 August OpenAI blocked a cluster of ChatGPT accounts it says very likely originated in Russia, with operators reaching the service through VPNs. The accounts had been writing social media comments that promoted the International Burke Institute, a website describing itself as a community of experts founded in 2025 and based in Israel.
OpenAI started by investigating AI-generated posts and found a larger operation behind them. The institute’s website was registered in February 2025, and the academic articles on it had been copied from real research, sometimes credited to people who had not written it. The site also carried a sovereignty index that ranked Russia favourably. Le Monde reported that it listed the UN and UNICEF among its partners, and named Mike Pompeo and Francis Fukuyama among its experts.
The operators wrote their prompts in Russian and asked ChatGPT for social media comments in English with any sign of a Russian author removed. The comments came in two forms: standalone posts pointing readers to IBI articles on X, LinkedIn, Facebook, Substack and Telegram, some of them from accounts carrying the institute’s name and logo, and replies left under posts by real Substack users, which usually asked them to follow the IBI channel.
Der Spiegel reported that it had found online traces connecting the institute to Dor Moriah, a charity run by Israelis with Russian and Ukrainian roots and with alleged ties to the Russian foreign ministry. Neither organisation answered questions. The accounts also fed Telegram channels, one of them with about 20,000 subscribers publishing material against the EU, the German government and Western support for Ukraine.
The Road to Ceuta: Mapping the Online Mobilisation Infrastructure
This summer’s Ceuta crossings set off one of the sharpest border debates the EU has seen in years: internal border checks returned, suspending Schengen was openly discussed, and 22 governments signed a letter criticising Spain’s migration policy.
It began weeks earlier on social media: Repsense analysed 25,335 posts and comments across nine platforms and found the same texts moving through the same fourteen Facebook groups in the same order, with gaps of 2 to 122 seconds, day after day. The groups were ordinary ones, where Moroccans abroad trade advice on visas, work and transport.
A handful of short texts carried the whole thing: a promise of how beautiful Europe is, a blessing for the road, jokes that the fence does not really hold anyone back, and a run of departure scenes telling anyone still hesitating that people like them had already gone.
Foreign-run pages are farming engagement with fake Pauline Hanson quotes
Facebook pages managed from Sri Lanka, the UK and the US put an invented line on AI images of John Farnham and Kylie Minogue, having them threaten to leave Australia if Hanson – the senator who leads One Nation, Australia’s right-wing populist party – became prime minister. The words belong to chef Neil Perry, who said them on 11 June. AAP says Hanson is a recurring subject for these pages because posts about her reliably draw comments and shares.
Murder For Hire Plot Exposes Latin American Disinformation Network
Fernando Cerimedo co-founded La Derecha Diario, advised Javier Milei, and runs digital campaigns for right-wing politicians across the region. Police arrested him over the shooting of his former partner, Nadia Beller, who survived. Searching properties linked to him, prosecutors found what they describe as a mini bot farm and more than half a million bolivianos in cash, along with dollars and euros. Mexican President Claudia Sheinbaum has accused the outlet of interfering in Mexico’s elections.
Japan is building a counter-disinformation unit inside its defence ministry
China has accused Japan of reviving militarism since Prime Minister Sanae Takaichi’s remarks in the Diet in November about a possible war over Taiwan. Tokyo will revise its three national security documents and raise defence spending by the end of the year, which is expected to draw more of the same. The new division, funded in the 2027 budget request, will decide when and how strongly to answer.
The Georgia Chamber of Commerce is launching a campaign to defend data centers
Georgia is second in the US for data centres, and county moratoriums are now live politics there. The chamber of commerce’s new alliance describes the backlash as “misinformation, disinformation, old information” and starts briefing policymakers in September. The Annenberg Public Policy Center found 61% of Americans opposed new data centres in their area this summer, up 12 points from the spring, with majorities in both parties against.
Last week we reported the more than 250% rise in anti-Ukrainian content in Polish after the Warsaw–Kyiv dispute opened.
Update: on 20 August Andriy Kovalenko, head of Ukraine’s Centre for Countering Disinformation, said Russia is expanding the effort through bot farms and new Polish-language networks on TikTok, and building Pravda-style sites in Polish, Ukrinform reported. Ukraine’s consulate general in Gdańsk reported an assault in Bydgoszcz on two Ukrainian teenagers, triggered by their speaking Ukrainian.
Get Narrative Ops in your inbox
Powered by Repsense narrative intelligence