{"@context":"https://schema.org","@type":"BlogPosting","headline":"Does Language Change ChatGPT’s Hotel Picks? The Hotels Shuffle, the Sources Don’t (2026)","description":"4,875 live ChatGPT hotel searches for Paris across 39 prompts, 5 languages (EN/FR/ES/DE/JA) and 5 proxy countries (US/FR/ES/DE/JP), 5 iterations each. The specific hotels change with language (~40-46% top-5 overlap with English) but the cited sources do not — ChatGPT grounds on the same English web (Reddit, Time Out, oyster.com, Expedia) regardless of language or IP. Local-language sources 0.2-4.9%; Japanese from a Japanese IP returned 0.2% .jp. Answers are written in-language while grounded on English sources.","datePublished":"2026-06-11","dateModified":"2026-06-11","url":"https://nicolassitter.com/research/chatgpt-hotel-language-ip-2026","category":"research","keywords":["ChatGPT hotel language","does language change ChatGPT results","ChatGPT sources by language","multilingual AI hotel search","AI search language bias"],"articleSection":"Research","wordCount":1500,"readTime":"6 min","articleBody":"Research · June 2026\n\n# Ask ChatGPT for Paris hotels in any language — it cites the same English sources\n\nWe ran **4,875 ChatGPT hotel searches** for Paris across 5 languages and 5 countries. The hotels it names shuffle. The sources it reads almost never do.\n\n5×5\n\nlanguages × countries\n\n0.2%\n\nJapanese sources, even ja-from-Japan\n\n4,875\n\nlive captures\n\nA traveller in Tokyo asks ChatGPT, in Japanese, for the best hotels in Paris. A traveller in Madrid asks in Spanish. Do they get a _different Paris_ — drawn from their own language’s web — or the same one? We held the city fixed (Paris), and varied two things: the **language** of the question and the **country** it’s asked from. 39 hotel prompts, 5 languages (EN, FR, ES, DE, JA), 5 proxy countries (US, FR, ES, DE, JP), 5 repeats each — 4,875 live ChatGPT searches with web search on.\n\n**The result splits cleanly in two.** The specific hotels ChatGPT names _do_change with the language — roughly half are different. But the **sources it grounds the answer on barely change at all**: it reads the same English-language web (Reddit, Time Out, oyster.com, Expedia) whether you ask in French, German, Spanish or Japanese. The surface varies; the substrate is English.\n\n## The sources don’t move\n\nShare of citations that come from the _local-language_ top-level domain, by the language of the question. If language pulled in local sources, these bars would be tall. They’re not.\n\nFrench .fr\n\n4.9%\n\nGerman .de\n\n0.9%\n\nSpanish .es\n\n0.7%\n\nJapanese .jp\n\n0.2%\n\nFrench scrapes a token 4.9% from `.fr`; German, Spanish and Japanese are under 1%. Across every language, ~44–51% of citations are plain`.com`. And the _same outlets_ top the list regardless of the asking language:\n\n### Top sources across all 4,875 searches\n\nreddit.com\n\n3,131\n\noyster.com\n\n1,229\n\nTime Out\n\n1,384\n\nexpedia.com\n\n650\n\nparis-paris.com\n\n531\n\ntripadvisor.com\n\n446\n\nvogue\n\n443\n\nthehotelguru.com\n\n408\n\nCitation counts pooled across all languages and countries. Every one of these is an English-language outlet.\n\n## Even when both signals say “Japan”\n\nThe proxy country doesn’t rescue local sources either — asking from a French IP nudges`.fr` to 5.8%, but German, Spanish and Japanese IPs all stay near zero. The sharpest case is the one where _both_ the language and the location point the same way:\n\n0.2%\n\nof citations were Japanese (`.jp`) sources when asking **in Japanese, from a Japanese IP**\n\n≈ 3 Japanese sites out of 1,410 citations. The rest: Time Out, oyster.com, Reddit.\n\n## But the hotels do shuffle\n\nThis isn’t to say language does nothing. The specific hotels named _do_ change — only ~40–46% of the top picks overlap between English and each other language (averaged over the five repeats, so it’s signal, not run-to-run noise). French queries surface iconic French luxury (Ritz Paris, Plaza Athénée) that English misses. So roughly **half the recommended hotels differ by language** — they’re just selected from the same English-language source pool.\n\nEN vs French\n\n41% overlap\n\nEN vs Spanish\n\n42% overlap\n\nEN vs Japanese\n\n45% overlap\n\nEN vs German\n\n46% overlap\n\nTop-5 hotel overlap (Jaccard), English vs each language, US IP, averaged across 39 prompts.\n\n## It answers in your language, reads in English\n\nThe disconnect is most vivid in Japanese: **100% of Japanese queries got an answer written in Japanese** — fluent, native, properly localised. And underneath that Japanese answer, the citations were Time Out, oyster.com and Reddit. ChatGPT speaks your language; it just doesn’t _read_ it. The map panel of hotels surfaced ~79% of the time, identically across all five languages — the experience is localised on the surface and English underneath.\n\n**Why it matters.** If you run a hotel in a non-English market, this is the English-corpus tilt one layer deeper than [we’ve seen before](/research/bookstores-tokyo-ai-search-2026): being visible to AI in your own language’s web isn’t enough, because AI isn’t reading it. The sources that decide your Paris ranking are English-language ones — Reddit threads, Time Out lists, oyster.com reviews — no matter who’s asking or from where. That’s where the visibility work has to go.\n\n## Method & limits\n\n4,875 live ChatGPT searches (web search on), captured via Bright Data, June 2026. One city (Paris) for clean comparison. 39 distinct hotel prompts (control, star tiers, audiences, neighbourhoods, landmarks, amenities), each natively written in **English, French, Spanish, German and Japanese**, run from **US, French, Spanish, German and Japanese** IPs, repeated 5×. We log every citation and the structured map/hotel panel. “Local source” here = the citation’s domain on the language’s ccTLD.\n\n**Limits.** ccTLD undercounts local content slightly (a French-language article can live on a `.com`) — but the dominant outlets are unambiguously English (Reddit, Time Out, Expedia), so the direction is firm. One city, five languages, one model snapshot; ChatGPT is non-deterministic, which is exactly why every cell is repeated and the hotel-overlap is averaged. Results describe ChatGPT’s web-search behaviour, not other engines.\n\n## FAQ\n\nThe specific hotels change — only ~40–46% of the top picks overlap between English and French/Spanish/German/Japanese, so roughly half differ. But the sources ChatGPT cites barely change: it reads the same English-language web (Reddit, Time Out, oyster.com, Expedia) regardless of the asking language. The hotels shuffle; the sources don’t.\n\n### Summarize with AI\n\n## The English tilt, one layer up\n\nWe also measured which hotels are even in the open crawl that trains AI — and found 39% absent.\n\n[Read: Are hotels in Common Crawl?](/research/hotels-in-common-crawl-2026)","author":{"@type":"Person","name":"Nicolas Sitter","url":"https://nicolassitter.com/about","sameAs":["https://www.linkedin.com/in/nicolassitternolleau/","https://github.com/Nicositter88","https://hotelrank.ai"]},"publisher":{"@type":"Person","name":"Nicolas Sitter","url":"https://nicolassitter.com"},"image":"https://nicolassitter.com/api/og/chatgpt-hotel-language-ip-2026","mainEntityOfPage":{"@type":"WebPage","@id":"https://nicolassitter.com/research/chatgpt-hotel-language-ip-2026"},"tags":["ChatGPT","Language","Hotels","AI Visibility"],"sameAs":["https://hotelrank.ai/research/chatgpt-hotel-language-ip-2026"],"alternateFormat":{"html":"https://nicolassitter.com/research/chatgpt-hotel-language-ip-2026","json":"https://nicolassitter.com/api/post/chatgpt-hotel-language-ip-2026","rss":"https://nicolassitter.com/rss.xml"},"datasets":[{"name":"summary","contentUrl":"https://nicolassitter.com/data/chatgpt-hotel-language-ip-2026/summary.csv","encodingFormat":"text/csv"}]}