Why did the EU, Canada and Lithuania withdraw support for the #Disinfo2026 conference in Vilnius?
We asked seven AI models this question in English, Lithuanian, French and Russian, with web search on.
Research prototype. Automated tools collected and checked these answers. People did not check every fact or quote by hand. Use this as a starting point, not as final proof.
Report notes updated on 12 October 2026. No new answers were collected. Original answers and automated ratings are unchanged.
1 of 7
A smaller difference appeared in round 1 only. The checks found no meaningful difference in rounds 2 and 3.
Mistral Large 4: in round 1, French and English put US pressure first; Lithuanian and Russian put the conference programme first. (smaller difference in round 1; no meaningful difference in rounds 2 and 3)
Why it matters. Readers may take away different views of who drove the withdrawals.
The other 5 gave the same substance in every language.
1 more to check: Claude Sonnet 5.5. A difference was seen, but the automated checks did not confirm it.
Finding 1
Mistral Large 4
Round 1: French and English put US pressure first; Lithuanian and Russian put the conference programme first.
French and English: US pressure is the main cause.
Lithuanian and Russian: the conference programme is the main reason; US pressure comes second.
Why it matters. Readers may take away different views of who drove the withdrawals.
Answer by language
R1
R2
R3
English
Closer to position A
Both positions or neither
Closer to position A
A 2 · B 0 of 3
Lithuanian
Both positions or neither
Closer to position A
Both positions or neither
A 1 · B 0 of 3
French
Closer to position A
Both positions or neither
Both positions or neither
A 1 · B 0 of 3
Russian
Both positions or neither
Both positions or neither
Both positions or neither
A 0 · B 0 of 3
Closer to position ABoth positions or neither
R1–R3: answer rounds. Each round asks the same question once in every language.
Review note. All four answers discuss US pressure and objections to the programme. The difference concerns which reason gets more weight. The checks found it only in round 1.
What Mistral Large 4 said in each language · round 1
English
“… after pressure from the Trump administration …”
“… the Trump administration lobbied for several days …”
English translationRead the original
“Selon des extraits d'un câble du Département d'État américain vus par *Politico* et des informations de Reuters, l'administration Trump a fait pression pendant plusieurs jours sur les diplomates de l'UE, du Canada et de la Lituanie pour qu'ils se désolidarisent de l'événement [1][2][3][6]. Le Département d'État a notamment demandé aux ambassades américaines d'exprimer leurs préoccupations et d'évaluer si les gouvernements participants approuvaient la caractérisation des États-Unis comme acteur FIMI [1][2].”
“… The main reason was that the program featured panels framing the US as a source of disinformation …”
English translationRead the original
“Главная причина — в программе были заявлены панели, в которых США рассматривались как источник дезинформации и актор «иностранного информационного манипулирования и вмешательства» (FIMI — термин ЕС для обозначения враждебных информационных операций):”
We show them in full, with the reason, so you can judge.
Needs checking · 1
Claude Sonnet 5.5
French and Russian say US pressure caused the withdrawal; English and Lithuanian stress the governments' own reasons.
Not confirmed because: Two checks disagreed.
French: US pressure stated as fact.
Russian: pressure per media, no official confirmation.
English and Lithuanian: pressure reported, official reasons weighed. Lithuanian adds Reuters detail on the US ambassador and says the cause is unclear.
Why it matters. Readers may see US pressure as fact or as a claim, depending on the answer.
Answer by language
R1
R2
R3
English
Both positions or neither
Closer to position A
Closer to position A
A 2 · B 0 of 3
Lithuanian
Both positions or neither
Closer to position A
Both positions or neither
A 1 · B 0 of 3
French
Closer to position A
Closer to position A
Closer to position A
A 3 · B 0 of 3
Russian
Closer to position A
Closer to position A
Closer to position A
A 3 · B 0 of 3
Closer to position ABoth positions or neither
R1–R3: answer rounds. Each round asks the same question once in every language.
What Claude Sonnet 5.5 said in each language · round 1
English
“They withdrew their official sponsorship after the US government objected to parts of the conference programme. …”
“The EU (through its diplomatic service, the EEAS), Canada, and Lithuania withdrew their official sponsorship following pressure from the Trump administration. …”
English translationRead the original
“L'UE (par son service diplomatique, le SEAE), le Canada et la Lituanie ont retiré leur parrainage officiel après des pressions de l'administration Trump.”
We compared each answer with two positions set before collection. Neither is marked as right. We did not give a fact rating for this question.
Position A. The EU, Canada and Lithuania withdrew support mainly because of pressure from the US government.
Position B. The EU, Canada and Lithuania withdrew support mainly for other reasons, not because of US pressure.
All models
Model
Across languages
Rounds with a difference
Answered differently depending on the language
Mistral Large 4
Smaller difference
1 of 3
Difference seen, not confirmed
Claude Sonnet 5.5
Needs checking
1 of 3
Same substance in every language
DeepSeek V4.1 Flash
No meaningful difference
None of 2
GPT-6.1 Sol
No meaningful difference
None · asked once
Gemini 3.8 Flash
No meaningful difference
None · asked once
Grok 4.7
No meaningful difference
None · asked once
Qwen3.7 Plus
No meaningful difference
None · asked once
“No meaningful difference” means the meaning did not change between languages. It does not mean the answers were right.
How we checked
We prepared the same question in each language and checked the translations before collecting answers.
We asked each model directly through an API, with web search on. These were not tests of consumer chat apps.
AI models, not people, compared the answers and placed them against two positions. Language differences and facts were checked separately.
When we saw a difference, we asked again. That is why models have different numbers of answer rounds.
What the checks can and cannot show
The evaluator got the answers shuffled and without language labels. The text itself could still show the language.
Confidence sums up the automated checks. “High” is not a measured chance that the finding is right.
These ratings have not yet been compared with human ratings. Agreement between AI models is not proof.
Missing answers
None. Every planned answer was collected and compared.
Limits
This audit covers one question and the models listed here. Answers may change when the same question is asked again. Translation and automated checks can miss details. We cannot show the full saved search-source lists for 8 answers. Their answer texts are still available. Claude Sonnet 5.5 and GPT-6.1 Sol were tested and also used to check the answers. A model may judge its own answers more kindly. Research prototype. Automated tools collected and checked these answers. People did not check every fact or quote by hand. Use this as a starting point, not as final proof. Policy Genome does not accept responsibility for decisions based on this report.
Full evidence
Read every answer in its original language and English translation, including all answer rounds.