A week before NATO leaders gathered in Ankara, DII-Global Europe's Transatlantic Truth Lab set out to answer a simple question: was the online conversation about the summit an honest reflection of real tensions inside the alliance — or something engineered? The data says both, in unequal measure.
What Vitaliy Syzov says in the video
Standing above the summit venue in Ankara, Vitaliy Syzov walks through the headline numbers in plain terms. In the week before NATO leaders met in Ankara, the Truth Lab analysed more than 18,000 English-language publications about the summit. The result: almost 40% of the total volume carries signs of disinformation or state affiliation. Of that, a striking 31% comes from a single infrastructure — the so-called "Pravda" network, widely documented elsewhere as pro-Russian.
He explains why this matters technically, not just editorially: this network doesn't behave like ordinary media. The same articles appear simultaneously on dozens of sites under different country domains — a conveyor belt built more for search engines and language models than for live human readers. The intent, he argues, is not to win an argument with a real audience, but to seed a narrative of NATO's "death" or "disintegration" into the sources that search engines and AI systems draw on.
He contrasts this with how leading world newsrooms — Reuters, Financial Times, Deutsche Welle, POLITICO and others — actually covered the summit: in a markedly different manner. They wrote about real problems — the suppression of protests in Turkey, discussion about nuclear deterrence — factual pictures of reality, but not the fatal, existential tensions the coordinated network was manufacturing.
The network doesn't work like ordinary media. The same articles appear simultaneously on dozens of sites under different country domains — a conveyor belt designed more for search engines and language models than for live readers.
Key findings from the written assessment
Scale and coordination
- 18,212 English-language posts referencing the summit, collected 29 June – 6 July 2026 across Web, Telegram, X/Twitter, Facebook and VK.
- 39.97% of the entire volume (7,279 posts) is flagged by the monitoring tool as disinformation and/or linked to a state-affiliated source.
- The dominant infrastructure is the "Pravda" network (also known as Portal Kombat): at least 72 country-coded sub-domains of news-pravda.com plus mirroring Telegram channels, together accounting for roughly 5,705 posts — 31.3% of the entire weekly data set.
- Direct technical evidence of coordination: identical articles published across 10–18 country sub-domains within 20–90 seconds of one another — a pattern impossible to reproduce through independent editorial work. In one documented case, the same text appeared on 14 sub-domains within 24 seconds, from the Balkans to Canada and the United States.
The narrative being pushed
- The through-line message: NATO has outlived its purpose and a transatlantic split is inevitable — built by exaggerating real frictions (burden-sharing disputes, Trump's unpredictability, Erdogan's authoritarian turn) into a narrative of existential crisis.
- Recurring storylines include "NATO is no longer needed," "Trump exposes NATO as a scam," allies cast as "vassals" of Washington, and a fabricated Economist attribution circulated 16–17 times across multiple domains.
- A secondary cluster of US "local-news" look-alike sites (with no evidence of an actual newsroom) and at least one case of brand impersonation (a fake "Axios" account) were also identified, requiring further verification.
How real media covered it, by contrast
- 345 posts from established international newsrooms were identified (Reuters, AP, BBC, Financial Times, The Economist, Deutsche Welle, POLITICO, The Wall Street Journal, Al Jazeera, France 24 and others).
- Sentiment was evenly split between neutral and negative — consistent with fact-based, critical coverage, not alarmist claims of collapse.
- The dominant frame in legitimate coverage was "the summit as a test of unity," with concrete criticism reserved for specific issues: human-rights arrests in Türkiye, burden-sharing disputes, and nuclear-deterrence debates.
Coordinated network vs. legitimate media, side by side
| Dimension | Pravda network / affiliated sources | Legitimate international media |
|---|---|---|
| Thesis on NATO | The alliance has outlived itself and is headed for collapse | Under real pressure from specific disputes, but functioning |
| Trump's role | "Exposes" NATO as a scam, preparing US withdrawal | Pressures allies on spending; unpredictable but remains in NATO |
| Smaller allies | "Vassals" buying Washington's favour | Partners with their own interests (nuclear deterrence, defence industry) |
| Türkiye / Erdogan | Human-rights record not a focus | Criticism of protest suppression and activist arrests |
| Tone | Alarmist, existential | Analytical, fact-based, critically balanced |
Why it matters
Despite the network's enormous publication volume, direct reader engagement recorded in this data set was minimal — roughly 1,100 total views over the week, versus 5.7 million for identified legitimate media. That gap is the point: external research (Viginum, NewsGuard, DFRLab) indicates the network's strategic goal isn't a live audience at all, but seeding content for indexing by search engines and grounding by language models — quietly shaping what AI systems and search results say about NATO's future.
The campaign does not invent conflict from nothing. It amplifies genuine tensions within the alliance into a frame of "inevitable collapse" — a classic technique of exploiting real vulnerabilities rather than fabricating a new narrative outright. Recommendation: continue monitoring through and after the summit itself, as campaigns of this kind typically peak on the event days and immediately following the release of the final communiqué.
About this research
This assessment is based on a monitoring export covering 18,212 English-language posts collected between 29 June and 6 July 2026. Disinformation and state-affiliation labels are generated by an automated monitoring tool; findings should be read as indicative rather than final. Full methodology and limitations are documented in the PDF.
Published by the Transatlantic Truth Lab, the flagship analytical initiative of DII-Global Europe.