[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-ai-chatbots-answer-the-same-moral-question-differently-by-language":10,"sections":39},{"siteName":4,"siteTagline":5,"publisherName":4,"contactEmail":6},"The Revision","Tech news, decoded.","editor@therevision.news",{"gaMeasurementId":8,"adsenseClientId":9},"G-ZW2MV82GYR","ca-pub-8533917693782264",{"article":11},{"id":12,"slug":13,"title":14,"dek":15,"body_md":16,"tags_json":17,"published_at":18,"created_at":19,"updated_at":20,"status":21,"review_note":22,"review_notes":23,"image_url":22,"persona_id":22,"persona_name":22,"section":30,"tags":31,"sources":34,"feedback":38,"feedback_at":22,"cost_usd":38,"total_tokens":38},6247,"ai-chatbots-answer-the-same-moral-question-differently-by-language","AI Chatbots Answer the Same Moral Question Differently by Language","AI models answer the same moral dilemma differently by language, and a new technique can steer and transfer chosen values across languages.","Ask an AI chatbot the same ethical question in English and Chinese, and you might get two different answers.\n\nResearchers built C-Voices, a dataset of 86,400 dilemma-based prompts across six languages, each pairing an action aligned with Chinese Social Values against a conflicting alternative. The 12 values span national, societal, and personal levels of Chinese culture. Testing several LLMs on identical dilemmas across languages, the team found that value preferences are model-dependent and language-sensitive - the same moral fork in the road produced divergent answers depending on which language it was asked in. To address this, the researchers also built a fine-tuning-free steering method that reads hidden-state differences between value-aligned and value-conflicting responses, then nudges a model's output toward a chosen value during inference.\n\nThe practical part is the transfer trick: a value vector extracted in one language can be applied to steer a model's behavior in a different language, without retraining. That means a company could calibrate a model's values once and expect the adjustment to travel across languages, a shortcut that sidesteps the inconsistency the researchers documented in the first place. The method also generalized to two existing value benchmarks, FLAMES and ValuePrism, suggesting the approach isn't just a one-dataset trick.\n\nTranslation, it turns out, isn't just a language problem for AI models - it's a values problem too.","[\"ai\",\"ai-alignment\",\"multilingual-ai\"]","2026-09-10T04:00:00.000Z","2026-09-10T13:15:29.046Z","2026-09-10T13:15:40.981Z","published",null,[24],{"id":25,"reviewer":26,"round":27,"reason":28,"status":29},"editor-r1","editor",1,"The dek claims researchers 'steered the outputs toward consistency,' but the body never describes cross-lingual consistency or transfer — it only says steering pushes output toward 'a target value' and lets a company 'dial' values after the fact, which is a different (and even contradictory) claim than achieving consistency, so either explain the cross-lingual value-vector transfer from the source in the body or rewrite the dek to match what the body actually supports.","resolved","ai",[30,32,33],"ai-alignment","multilingual-ai",[35],{"name":36,"url":37},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2609.08515",0,{"sections":40},[41,45,49,54,59,64,69,73,78,83,88,93,98,103],{"name":42,"slug":30,"count":43,"latest_published_at":44},"AI",3480,"2026-09-11T04:00:00.000Z",{"name":46,"slug":47,"count":48,"latest_published_at":44},"Security","security",628,{"name":50,"slug":51,"count":52,"latest_published_at":53},"Policy","policy",336,"2026-09-11T00:56:21.000Z",{"name":55,"slug":56,"count":57,"latest_published_at":58},"Deals","deals",179,"2026-06-29T20:02:07.000Z",{"name":60,"slug":61,"count":62,"latest_published_at":63},"Hardware","hardware",153,"2026-09-09T15:12:32.000Z",{"name":65,"slug":66,"count":67,"latest_published_at":68},"Consumer Tech","consumer-tech",99,"2026-09-09T17:27:33.000Z",{"name":70,"slug":71,"count":72,"latest_published_at":44},"Science","science",98,{"name":74,"slug":75,"count":76,"latest_published_at":77},"Software","software",75,"2026-09-10T20:41:21.000Z",{"name":79,"slug":80,"count":81,"latest_published_at":82},"Dev Tools","dev-tools",69,"2026-08-18T04:00:00.000Z",{"name":84,"slug":85,"count":86,"latest_published_at":87},"Startups","startups",55,"2026-09-09T23:14:29.000Z",{"name":89,"slug":90,"count":91,"latest_published_at":92},"Gaming","gaming",43,"2026-09-10T12:18:06.000Z",{"name":94,"slug":95,"count":96,"latest_published_at":97},"General","general",41,"2026-09-08T01:57:23.000Z",{"name":99,"slug":100,"count":101,"latest_published_at":102},"Reviews","reviews",20,"2026-06-24T12:00:01.000Z",{"name":104,"slug":105,"count":106,"latest_published_at":107},"How-To","how-to",6,"2026-06-16T09:00:00.000Z"]