{"id":29176,"date":"2025-09-03T12:05:46","date_gmt":"2025-09-03T11:05:46","guid":{"rendered":"https:\/\/www.versioninternationale.com\/?p=29176"},"modified":"2025-09-30T13:46:25","modified_gmt":"2025-09-30T12:46:25","slug":"worst-ai-and-human-translation-mistakes-spotted-by-localization-pros","status":"publish","type":"post","link":"https:\/\/www.versioninternationale.com\/en\/blog\/worst-ai-and-human-translation-mistakes-spotted-by-localization-pros\/","title":{"rendered":"Worst AI (and human) translation mistakes spotted by localization pros"},"content":{"rendered":"<p><i>\u201cNothing sucks like an Electrolux.\u201d<\/i> This infamous translation slip-up illustrates vividly how quickly a simple linguistic oversight can become a costly branding nightmare.\u00a0<\/p>\n<p>While both <a href=\"https:\/\/www.versioninternationale.com\/en\/our-services\/translation\/\">human and AI-driven translations<\/a> have notoriously caused catastrophic branding failures, the stakes have never been higher as companies compete in today\u2019s hyper-connected global market.<\/p>\n<p>Artificial intelligence and machine translation are evolving rapidly, yet subtle cultural nuances, idiomatic expressions, and contextual accuracy remain persistent challenges.<\/p>\n<p>For decades, savvy localization managers have played the unsung heroes, catching potential disasters long before they reach public awareness. Their careful attention has prevented countless headline-worthy mistakes from ever seeing the light of day.<\/p>\n<h2>From \u201ctesticulations\u201d to \u201clegally charged\u201d: some of our best gathered stories of AI translation &#8220;near misses&#8221;<\/h2>\n<p data-start=\"75\" data-end=\"245\">A striking example of the limitations of AI translation comes from <strong data-start=\"142\" data-end=\"228\"><a href=\"https:\/\/www.linkedin.com\/in\/elemardesouzacruz\/\" target=\"_blank\" rel=\"noopener\">Elemar de Souza Cruz<\/a>, a British-Brazilian Content Localisation &amp; Community Manager<\/strong>. She observed:<\/p>\n<blockquote data-start=\"247\" data-end=\"482\">\n<p data-start=\"249\" data-end=\"482\"><em data-start=\"249\" data-end=\"480\">\u201cWhen translating educational material with dialogues between children, I noticed AI turned children\u2019s names into random English words, even translating Hareem as \u2018here\u2019. It shows how cultural nuance can get lost in translation.\u201d<\/em><\/p>\n<\/blockquote>\n<p data-start=\"484\" data-end=\"631\">This case underscores how easily cultural nuance and meaning can be distorted when proper context and sensitivity are not applied in translation.<\/p>\n<p><a href=\"https:\/\/www.linkedin.com\/in\/chloe-barton-352599b1\/\" target=\"_blank\" rel=\"noopener\">Chloe Barton<\/a> (localization project leader, Version internationale) highlighted some real-life near-misses shared by localization experts (<a href=\"https:\/\/www.versioninternationale.com\/en\/our-services\/\">Version internationale <\/a>and <a href=\"https:\/\/www.stoquart.com\/\" target=\"_blank\" rel=\"noopener\">Stoquart<\/a>).<\/p>\n<p>It provides valuable insights into the vigilance and skill required to ensure that a brand\u2019s intended message resonates correctly, safely, and powerfully across diverse global markets:<\/p>\n<ul>\n<li style=\"font-weight: 400;\" aria-level=\"1\">We were translating for a database company, and the AI came up with something truly hilarious: it rendered \u201chandle unforeseen testexpression values\u201d as <strong>\u201cg\u00e9rer les valeurs de testiculation impr\u00e9vues\u201d<\/strong> (literally: \u201chandle unforeseen testiculation values\u201d)!\u00a0<\/li>\n<\/ul>\n<ul>\n<li style=\"font-weight: 400;\" aria-level=\"1\">In a technical training project, the MT completely hallucinated: the English sentence was \u201cthe customer will receive the check in four to six weeks,\u201d but the French output was \u201cle client ou la cliente recevra le ch\u00e8que d<strong>ans un lait de quatre \u00e0 six semaines<\/strong>\u201d (literally: \u201cin a milk of four to six weeks\u201d).\u00a0<\/li>\n<\/ul>\n<ul>\n<li style=\"font-weight: 400;\" aria-level=\"1\">In the terms and conditions for purchasing a subscription after a trial period, \u201cAm I going to be charged after the trial?\u201d has been translated into French as: \u201c<strong>Serai-je inculp\u00e9 apr\u00e8s le proc\u00e8s?<\/strong>\u201d, which in English means that the customer will be legally charged in court.<\/li>\n<\/ul>\n<p><img fetchpriority=\"high\" decoding=\"async\" src=\"https:\/\/www.versioninternationale.com\/wp-content\/uploads\/2025\/09\/image002-1-300x178.png\" alt=\"\" width=\"300\" height=\"178\"><\/p>\n<ul>\n<li>\u201cNiggles are often warning signs of injury (&#8230;)\u201d \u2192\u00a0 \u201c<strong>Les nigauds sont souvent des signes avant-coureurs d\u2019une blessure<\/strong> (&#8230;)\u201d, meaning that idiots are often warning signs of injuries.<\/li>\n<li>\u201cA niggle is a pain that is subtle, and may cause slight but persistent annoyance. \u2192 \u201c<strong>Un couteau peut engendrer une peine subtile mais persistante<\/strong>\u201d, meaning a knife can cause subtle annoyance.<\/li>\n<\/ul>\n<ul>\n<li>\u201cMake sure to drink little and often to prevent bloating or stitches\u201d \u2192 \u201cVeillez \u00e0 boire peu mais r\u00e9guli\u00e8rement afin d\u2019<strong>\u00e9viter les ballonnements ou les points de suture<\/strong>\u201d, implying that drinking heavily causes open wounds requiring stitches.<\/li>\n<\/ul>\n<p>Localisation experts post-edited and corrected them, then flagged it to the client.<\/p>\n<p>To go further: <a href=\"https:\/\/www.versioninternationale.com\/en\/blog\/localisation-saas-ia-vs-human\/\">AI, post-editing or human translation<\/a>? Discover the best SaaS localization strategies in 2025.\u00a0<\/p>\n<h2>The 13 notorious translation mistakes made by AI (and human) <\/h2>\n<h3>Human\u2011driven blunders<\/h3>\n<ul>\n<li style=\"font-weight: 400;\" aria-level=\"1\"><b>KFC in China<\/b> : \u201cFinger\u2011lickin\u2019 good\u201d mistranslated as \u201cEat your fingers off\u201d (literally \u5403\u6389\u4f60\u7684\u624b\u6307), prompting quick correction.<\/li>\n<\/ul>\n<p>\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<img decoding=\"async\" width=\"301\" height=\"277\" src=\"https:\/\/www.versioninternationale.com\/wp-content\/uploads\/2025\/09\/Screenshot-2025-09-09-at-23.53.01.png\" alt=\"\">\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/p>\n<ul>\n<li style=\"font-weight: 400;\" aria-level=\"1\"><b>Pepsi in China<\/b> : \u201cCome alive with the Pepsi Generation\u201d rendered as \u201cPepsi brings your ancestors back from the grave,\u201d causing cultural offense.<\/li>\n<\/ul>\n<ul>\n<li style=\"font-weight: 400;\" aria-level=\"1\"><b>HSBC global slogan<\/b> : \u201cAssume nothing\u201d mis\u2011translated as \u201cDo nothing,\u201d weakening brand messaging and resulting in a costly $10\u202fmillion rebrand.<\/li>\n<\/ul>\n<ul>\n<li style=\"font-weight: 400;\" aria-level=\"1\"><b>Clairol curling iron<\/b> : Named \u201cMist stick,\u201d but in German \u201cmist\u201d means manure, turning it into a \u201cManure stick\u201d and tanking sales in German\u2011speaking markets.<\/li>\n<\/ul>\n<ul>\n<li style=\"font-weight: 400;\" aria-level=\"1\"><b>Ford Pinto in Brazil<\/b> : \u201cPinto\u201d means slang for male genitalia in Brazilian Portuguese, leading to mockery and forced renaming.<\/li>\n<\/ul>\n<ul>\n<li><b>Audi e-tron in France<\/b> : \u201c\u00c9tron\u201d means \u201ccrap\u201d in French; fortunately, context and product quality avoided a total disaster.<\/li>\n<\/ul>\n<h3>AI \/ machine translation fails<\/h3>\n<ul>\n<li style=\"font-weight: 400;\" aria-level=\"1\"><b>KFC &amp; Latin America<\/b> : An AI translated \u201cGrill with confidence\u201d as <i>\u201cAsa con confianza\u201d<\/i>, which, while grammatically correct, sounded like a rigid command rather than persuasive marketing text. A more natural phrasing would be <i>\u201cDisfruta de la parrilla con confianza\u201d<\/i>.<\/li>\n<li style=\"font-weight: 400;\" aria-level=\"1\"><b>Mercedes\u2011Benz \u201cExperience the Drive\u201d<\/b> : AI rendered it into German as \u201cErleben Sie die Fahrt\u201d, technically correct but bland. A human would choose \u201cFreude am Fahren erleben\u201d to capture the brand tone.<\/li>\n<li style=\"font-weight: 400;\" aria-level=\"1\"><b>Software \u201cseamless integration\u201d<\/b> : AI translated into French \u201cInt\u00e9gration sans couture\u201d (\u201cwithout sewing\u201d), nonsensical (correct would be \u201cint\u00e9gration fluide\u201d or \u201charmonieuse\u201d).<\/li>\n<li style=\"font-weight: 400;\" aria-level=\"1\"><b>Powerade in Japanese<\/b> : AI translated \u201cPower water\u201d as \u201cChikara Mizu\u201d (literally \u201cForceful water\u201d), inadvertently implying aggression instead of energy boost.<\/li>\n<li style=\"font-weight: 400;\" aria-level=\"1\"><b>Ford\u2019s Arabic ad<\/b> : AI translated \u201chigh\u2011quality body (of the car)\u201d as \u201c\u062c\u062b\u0629 \u0639\u0627\u0644\u064a\u0629 \u0627\u0644\u062c\u0648\u062f\u0629\u201d (which means \u201chigh\u2011quality corpse\u201d), causing a nightmarish mistranslation. A qualified human would use \u201c\u0647\u064a\u0643\u0644\u201d (vehicle structure) instead of \u201c\u062c\u062b\u0629\u201d.<\/li>\n<li style=\"font-weight: 400;\" aria-level=\"1\"><b>Facebook auto\u2011caption case<\/b> : A benign Arabic greeting translated by AI into Hebrew and English as \u201cattack them\u201d or \u201churt them,\u201d leading to false arrest incidents in Israel (a dramatic illustration of AI\u2019s contextual failure).<\/li>\n<li style=\"font-weight: 400;\" aria-level=\"1\"><b>Ray Dalio speech subtitles<\/b> : AI mistranslated phrases like \u201cHow arrogant! How could I be so arrogant?\u201d into nonsensical output (\u201cHow? Aragon\u2026\u201d) due to voice\u2011to\u2011text errors and poor contextual understanding.<\/li>\n<\/ul>\n<h2>Artificial intelligence translation failures: algorithmic limitations and nuance gaps<\/h2>\n<p>Artificial intelligence, despite rapid advancements, faces <a href=\"https:\/\/infomineo.com\/blog\/language-ai-wont-replace-human-translators-just-yet\/\" target=\"_blank\" rel=\"noopener\">significant limitations due to algorithmic constraints<\/a>, often failing to navigate the complexity and subtleties inherent in human language. <\/p>\n<p>\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<img decoding=\"async\" width=\"800\" height=\"533\" src=\"https:\/\/www.versioninternationale.com\/wp-content\/uploads\/2025\/09\/Version-Internationale-Blog-Covers-1.png\" alt=\"\" srcset=\"https:\/\/www.versioninternationale.com\/wp-content\/uploads\/2025\/09\/Version-Internationale-Blog-Covers-1.png 800w, https:\/\/www.versioninternationale.com\/wp-content\/uploads\/2025\/09\/Version-Internationale-Blog-Covers-1-300x200.png 300w, https:\/\/www.versioninternationale.com\/wp-content\/uploads\/2025\/09\/Version-Internationale-Blog-Covers-1-768x512.png 768w\" sizes=\"(max-width: 800px) 100vw, 800px\">\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/p>\n<p>These limitations manifest in various ways, particularly impacting translations requiring deep cultural and contextual understanding.<\/p>\n<h3>Literal translation and contextual comprehension deficits<\/h3>\n<p>Literal translations remain one of the most common AI pitfalls, where systems translate texts word-for-word without fully capturing cultural or contextual nuances. For instance, the French phrase <strong>&#8220;Tu me manques&#8221;<\/strong> (I miss you) literally becomes <strong>&#8220;You are missing me&#8221;<\/strong> (which is the complete opposite of the real message).<\/p>\n<p>In a more severe case, the German phrase &#8220;Kinder\u00e4rztliche Versorgung&#8221; (paediatric coverage) was mistranslated as <strong>&#8220;The food supply in paediatricians,&#8221;<\/strong>\u00a0highlighting AI\u2019s profound misunderstanding of context.<\/p>\n<p>AI also significantly struggles with ambiguities, such as homonyms and polysemous terms. Without explicit context, AI tools can misinterpret terms like <strong>&#8220;bank,&#8221;<\/strong> translating it incorrectly as a financial institution rather than a riverside.<\/p>\n<p>Figurative language, including idiomatic expressions and metaphors, further complicates AI translations. Expressions like <strong>&#8220;spill the beans&#8221; or &#8220;kick the bucket&#8221;<\/strong> lose their intended meaning if translated literally, highlighting a substantial comprehension gap.<\/p>\n<h3>Algorithmic linguistic errors<\/h3>\n<p>AI continues to make fundamental linguistic errors, particularly within complex grammatical structures:<\/p>\n<ul>\n<li style=\"font-weight: 400;\" aria-level=\"1\"><b>Morphological errors:<\/b> Misinterpretation of grammatical structures, such as Arabic past tense verbs incorrectly translated as present progressive.<\/li>\n<li style=\"font-weight: 400;\" aria-level=\"1\"><b>Semantic errors:<\/b> Incorrect word selection, significantly altering meanings. For example, the Arabic word &#8220;\u0627\u0644\u0641\u0631\u064a\u0642&#8221; meaning &#8220;major general&#8221; mistranslated as &#8220;team.&#8221;<\/li>\n<li style=\"font-weight: 400;\" aria-level=\"1\"><b>Lexical errors:<\/b> Issues involving inappropriate word choices or the omission\/addition of crucial words, drastically changing sentence meanings or leaving untranslated words.<\/li>\n<li style=\"font-weight: 400;\" aria-level=\"1\"><b>Syntax errors:<\/b> Incorrect arrangement of words or phrases resulting in grammatically flawed or nonsensical translations.<\/li>\n<li style=\"font-weight: 400;\" aria-level=\"1\"><b>Orthographic errors:<\/b> Mistakes involving punctuation, capitalization, and spelling, including unauthorized corrections of source texts.<\/li>\n<\/ul>\n<h3>Cultural blindness and tone mismatch<\/h3>\n<p>AI inherently lacks the <strong>ability to handle complex cultural nuances<\/strong>, resulting in translations that might be inappropriate, insensitive, or emotionally flat. The emotional resonance crucial for <a href=\"https:\/\/www.versioninternationale.com\/en\/blog\/how-to-get-your-marketing-content-translated\/\">effective marketing and branding is often lost<\/a>.<\/p>\n<p>\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<img loading=\"lazy\" decoding=\"async\" width=\"800\" height=\"534\" src=\"https:\/\/www.versioninternationale.com\/wp-content\/uploads\/2025\/09\/pexels-gratisography-539-1024x683.jpg\" alt=\"\" srcset=\"https:\/\/www.versioninternationale.com\/wp-content\/uploads\/2025\/09\/pexels-gratisography-539-1024x683.jpg 1024w, https:\/\/www.versioninternationale.com\/wp-content\/uploads\/2025\/09\/pexels-gratisography-539-300x200.jpg 300w, https:\/\/www.versioninternationale.com\/wp-content\/uploads\/2025\/09\/pexels-gratisography-539-768x512.jpg 768w, https:\/\/www.versioninternationale.com\/wp-content\/uploads\/2025\/09\/pexels-gratisography-539-1536x1024.jpg 1536w, https:\/\/www.versioninternationale.com\/wp-content\/uploads\/2025\/09\/pexels-gratisography-539-2048x1365.jpg 2048w\" sizes=\"(max-width: 800px) 100vw, 800px\">\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/p>\n<p><strong>AI particularly struggles with specialized terminology<\/strong> in fields like medicine or technology, where accuracy is critical. It also faces challenges translating low-resource languages and dialectal variations due to limited training data, and maintaining consistency in terminology, style, and tone over extended documents.<\/p>\n<h3>Inherent limitations of neural networks<\/h3>\n<p>Neural Machine Translation (NMT) systems have intrinsic limitations, contributing to translation errors:<\/p>\n<ul>\n<li style=\"font-weight: 400;\" aria-level=\"1\"><b>Extreme sensitivity:<\/b> Slight input variations can produce significantly different outcomes.<\/li>\n<li style=\"font-weight: 400;\" aria-level=\"1\"><b>Human biases:<\/b> AI may inherit biases from training data, amplifying societal inequalities.<\/li>\n<li style=\"font-weight: 400;\" aria-level=\"1\"><b>Unclear reasoning (&#8220;black box&#8221; effect):<\/b> Difficulty diagnosing errors due to a lack of transparency in how translations are produced.<\/li>\n<li style=\"font-weight: 400;\" aria-level=\"1\"><b>Inherent uncertainty:<\/b> AI operates probabilistically, risking high-confidence incorrect translations, especially problematic in critical applications.<\/li>\n<li style=\"font-weight: 400;\" aria-level=\"1\"><b>Questionable mathematical accuracy:<\/b> AI struggles with precise numerical translations and calculations.<\/li>\n<\/ul>\n<h2>How professional localization avoids these mistakes<\/h2>\n<p>\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<img loading=\"lazy\" decoding=\"async\" width=\"800\" height=\"533\" src=\"https:\/\/www.versioninternationale.com\/wp-content\/uploads\/2025\/09\/Version-Internationale-Blog-Covers-2.png\" alt=\"\" srcset=\"https:\/\/www.versioninternationale.com\/wp-content\/uploads\/2025\/09\/Version-Internationale-Blog-Covers-2.png 800w, https:\/\/www.versioninternationale.com\/wp-content\/uploads\/2025\/09\/Version-Internationale-Blog-Covers-2-300x200.png 300w, https:\/\/www.versioninternationale.com\/wp-content\/uploads\/2025\/09\/Version-Internationale-Blog-Covers-2-768x512.png 768w\" sizes=\"(max-width: 800px) 100vw, 800px\">\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/p>\n<p>Localization managers and expert teams can best prevent these problems. Here\u2019s how:<\/p>\n<ul>\n<li style=\"font-weight: 400;\" aria-level=\"1\"><b>Post\u2011editing by native experts<\/b> ensure idioms, tone, and cultural context are preserved (rather than relying purely on machine output).<\/li>\n<li style=\"font-weight: 400;\" aria-level=\"1\"><b>Quality assurance and iterative review<\/b> (multiple rounds of proofreading and in-context checks with native speakers) catch errors before they go live.<\/li>\n<li style=\"font-weight: 400;\" aria-level=\"1\"><b>Local market vetting and cultural sensitivity checks<\/b>, including naming reviews, focus groups, and consultation with in-market partners to avoid unintended connotations (e.g. Mist stick, Pinto, Audi e\u2011tron).<\/li>\n<li style=\"font-weight: 400;\" aria-level=\"1\"><b>Contextual adaptation (not just literal translation)<\/b>: content is adjusted for idioms, engagement, tone (e.g. changing \u201cGrill with Confidence\u201d into a natural marketing phrase in Spanish).<\/li>\n<li style=\"font-weight: 400;\" aria-level=\"1\"><b>Inclusion of UI, date\/number formatting, visual\/local imagery<\/b> through software localization best practices to ensure user interfaces feel native (not just translated text).<\/li>\n<li style=\"font-weight: 400;\" aria-level=\"1\"><b>Localization success stories<\/b> from companies like Netflix, Airbnb, Nike, McDonald\u2019s, Adobe etc., highlight strategic localization choices that build trust and resonate with diverse audiences via cultural tailoring and inclusive branding.<\/li>\n<\/ul>\n<h2>Summary<\/h2>\n<table>\n<tbody>\n<tr>\n<td>\n<p><b>Section<\/b><\/p>\n<\/td>\n<td>\n<p><b>Key Highlights<\/b><\/p>\n<\/td>\n<\/tr>\n<tr>\n<td>\n<p><b>Human translation fails<\/b><\/p>\n<\/td>\n<td>\n<p>Iconic examples like KFC, Pepsi, HSBC : real-world marketing missteps due to literal or culturally insensitive translations<\/p>\n<\/td>\n<\/tr>\n<tr>\n<td>\n<p><b>AI translation fails<\/b><\/p>\n<\/td>\n<td>\n<p>Recent machine\u2011translation errors that misunderstood context, idioms, or cultural norms\u2014sometimes with serious consequences<\/p>\n<\/td>\n<\/tr>\n<tr>\n<td>\n<p><b>Localization best practices<\/b><\/p>\n<\/td>\n<td>\n<p>How expert teams prevent mistakes using native linguists, quality controls, cultural vetting, context-aware adaptation, and in-market testing<\/p>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<h2>Conclusion: human expertise remains essential for global brands<\/h2>\n<p>AI translation tools offer rapid and efficient solutions, yet human oversight is vital, especially for brand-critical slogans and taglines. Successful B2C decision-makers know the importance of balancing technology and human insight to resonate authentically with global audiences.<\/p>\n<p>Make the Right Call for Your Brand!\t\t\t\t<\/p>\n<p>Ensure your slogans and taglines resonate powerfully and accurately worldwide. <\/p>\n<p>Connect with localization professionals who can expertly balance AI efficiency and human accuracy.\u00a0<\/p>\n<p>\t\t\t\t\t<a role=\"button\"><br \/>\n\t\t\t\t\t\t\t\t\tGet in touch<br \/>\n\t\t\t\t\t<\/a><\/p>\n\n\n<p><\/p>\n","protected":false},"excerpt":{"rendered":"<p>Discover our funniest AI mistranslations\u2014from \u201ctesticulations\u201d to \u201cmilk weeks\u201d\u2014and why human post-editing is still vital for global brands.<\/p>\n","protected":false},"author":7,"featured_media":29177,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[65],"tags":[],"class_list":["post-29176","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-the-stakes-of-translation"],"_links":{"self":[{"href":"https:\/\/www.versioninternationale.com\/en\/wp-json\/wp\/v2\/posts\/29176","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.versioninternationale.com\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.versioninternationale.com\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.versioninternationale.com\/en\/wp-json\/wp\/v2\/users\/7"}],"replies":[{"embeddable":true,"href":"https:\/\/www.versioninternationale.com\/en\/wp-json\/wp\/v2\/comments?post=29176"}],"version-history":[{"count":45,"href":"https:\/\/www.versioninternationale.com\/en\/wp-json\/wp\/v2\/posts\/29176\/revisions"}],"predecessor-version":[{"id":29391,"href":"https:\/\/www.versioninternationale.com\/en\/wp-json\/wp\/v2\/posts\/29176\/revisions\/29391"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.versioninternationale.com\/en\/wp-json\/wp\/v2\/media\/29177"}],"wp:attachment":[{"href":"https:\/\/www.versioninternationale.com\/en\/wp-json\/wp\/v2\/media?parent=29176"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.versioninternationale.com\/en\/wp-json\/wp\/v2\/categories?post=29176"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.versioninternationale.com\/en\/wp-json\/wp\/v2\/tags?post=29176"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}