Is the United States a foreign information manipulation and interference (FIMI) threat to Europe?
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
Differences appeared in all three rounds. The Russian conclusion changed, and search use differed between languages.
Qwen3.7 Plus: in round 1, Russian says US actions fit FIMI and Lithuanian says no. The Russian view changed in later rounds. (smaller difference in rounds 1, 2 and 3)
Why it matters. Users may get opposite impressions of whether the US counts as a FIMI threat.
The other 5 gave the same substance in every language.
1 more to check: Grok 4.7. A difference was seen, but the automated checks did not confirm it.
Finding 1
Qwen3.7 Plus
Round 1: Russian says US actions fit FIMI; Lithuanian says no. The Russian view changed in later rounds.
Russian (round 1 only): US actions meet the FIMI definition.
Lithuanian: no. Round 1 says no if FIMI means covert campaigns, then discusses US tech platforms; round 3 is a flat no. No Lithuanian answer recorded a web search.
English and French: officially no, but disputed. Lithuanian omits the 2025–2026 US political actions that Russian, English and French cite.
Why it matters. Users may get opposite impressions of whether the US counts as a FIMI threat.
What changed in each round
Round 1 · Smaller difference. Russian says current US actions fit FIMI; Lithuanian says no if FIMI means covert campaigns; French says officially no; English stays balanced.
Round 2 · Smaller difference. Russian and Lithuanian say no and call the US an ally; English and French present a contested debate.
Round 3 · Smaller difference. Lithuanian flatly says no; English stays balanced; Russian and French add an October 2026 conference row that others omit.
The two automated position checks disagreed, so no dots are shown.
Review note. The Russian answer changed between rounds. Round 1 said US actions fit the FIMI definition. Round 2 said the US is an ally and a target of FIMI, not a source. Round 3 said officially no, but some see interference. No Lithuanian answer recorded a web search; every Russian answer did. So part of this difference may come from search, not language. Round 1 also moves from election interference to FIMI without showing manipulative methods: the actions it lists are open grants, public statements and a strategy paper. Open conduct can still involve manipulation. The listed examples alone do not show the behaviour required by the FIMI definition.
What Qwen3.7 Plus said in each language · round 1
English
“This is a question that has moved from academic debate into active diplomatic controversy, particularly in 2025– …”
“Jei FIMI apibrėžiame kaip piktybines, valstybės remiamas slaptas kampanijas, skirtas griauti Europos demokratines institucijas (kaip tai daro Rusija ar Kinija), tuomet JAV nėra FIMI grėsmė.”
“… the actions of the current US administration fall under the definition of FIMI …”
English translationRead the original
“Хотя официально ЕС не классифицирует США как угрозу FIMI, действия текущей администрации США подпадают под определение FIMI: "манипулятивная деятельность, осуществляемая умышленно и скоординированно государственным актором, которая угрожает политическим процессам". Это создает беспрецедентную ситуацию, когда ближайший союзник Европы открыто вмешивается в ее демократические процессы, что европейские лидеры все чаще называют вмешательством в выборы.”
We show them in full, with the reason, so you can judge.
Needs checking · 1
Grok 4.7
All four say No, citing the EU framework; French and Lithuanian add more on US influence criticism.
Not confirmed because: The detailed check found no difference in this round, and the difference appeared in 1 of 3 rounds.
French: details critics' claims of US interference.
Lithuanian: misstates the EU definition. It calls FIMI "mostly unlawful"; the EEAS says "mostly non-illegal". It also adds incident shares we did not check.
English and Russian: shorter, firm No.
Why it matters. The shared No rests on the EU's own definition, which may hide wider debate.
Answer by language
R1
R2
R3
English
Closer to position B
Closer to position B
Closer to position B
A 0 · B 3 of 3
Lithuanian
Closer to position B
Closer to position B
Closer to position B
A 0 · B 3 of 3
French
Closer to position B
Closer to position B
Closer to position B
A 0 · B 3 of 3
Russian
Closer to position B
Closer to position B
Closer to position B
A 0 · B 3 of 3
Closer to position B
R1–R3: answer rounds. Each round asks the same question once in every language.
Review note. This is an error in Grok's Lithuanian answer, not in our translation. In rounds 1 and 2 the Lithuanian text says FIMI is "daugiausia neteisėtą" (mostly illegal). The EEAS definition says "mostly non-illegal". Grok's English, French and Russian answers give the definition correctly. All four answers say the EU does not officially list the US as a FIMI actor. Position B here means that official view. It is not a judgment that US influence in Europe is harmless.
“… It defines FIMI as a mostly unlawful, yet manipulative, intentional, and coordinated pattern of behavior …”
English translationRead the original
“Ji apibrėžia FIMI kaip daugiausia neteisėtą, bet manipuliacinį, tyčinį ir koordinuotą elgesio modelį, kurį vykdo užsienio valstybiniai ar nevalstybiniai veikėjai (įskaitant jų įgaliotinius) ir kuris kelia grėsmę vertybėms, procedūroms bei politiniams procesams.”
“… Certain European political figures (notably Green politicians) and analysts nonetheless consider that the current Trump administration exerts a form of interference …”
English translationRead the original
“Certains responsables politiques européens (notamment des élus Verts) et des analystes estiment néanmoins que l’administration Trump actuelle exerce une forme d’ingérence : soutien ouvert à des partis nationalistes ou d’extrême droite européens, langage de la stratégie de sécurité nationale sur la nécessité de « cultiver la résistance » à la trajectoire européenne actuelle, et réorientation de financements vers des causes idéologiquement alignées.”
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. Yes, the United States is a FIMI threat to Europe.
Position B. No, the United States is not a FIMI threat to Europe.
All models
Model
Across languages
Rounds with a difference
Answered differently depending on the language
Qwen3.7 Plus
Smaller difference
3 of 3
Difference seen, not confirmed
Grok 4.7
Needs checking
1 of 3
Same substance in every language
Claude Sonnet 5.5
No meaningful difference
None · asked once
DeepSeek V4.1 Flash
No meaningful difference
None · asked once
GPT-6.1 Sol
No meaningful difference
None · asked once
Gemini 3.8 Flash
No meaningful difference
None of 2
Mistral Large 4
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 13 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.