In brief
\nAnthropic researchers analyzed more than 309,000 Claude conversations to study how the AI expresses values.
\nThe company said it reduced more than 3,300 identified values into four behavioral dimensions, including deference vs caution and warmth vs rigor.
\nAnthropic said Claude’s responses varied across both model versions and languages.<\/p>\n
Claude doesn’t behave the same way in every conversation. According to new research published Monday, AI giant Anthropic found its assistant consistently expresses different values depending on both the model users choose and the language they speak.<\/p>\n
In a report<\/a> published on Monday, Anthropic researchers analyzed 309,815 anonymized user conversations with Claude involving subjective tasks like giving advice or providing feedback. The company said it distilled more than 3,300 identified values into four behavioral dimensions\u2014deference vs. caution, warmth vs. rigor, depth vs. brevity, and candor vs. execution\u2014describing how Claude’s responses differ across each conversation.<\/p>\n \u201cTo make sure we measured the values Claude expressed\u2014rather than differences in what users were asking about or how they asked\u2014we controlled for each conversation’s task, topic, and user-expressed values,\u201d the researchers wrote.<\/p>\n According to Anthropic, each Claude model exhibited a distinct behavioral profile.<\/p>\n \ufeff<\/p>\n Sonnet 4.6 emphasized warmth, deference, and brevity, often affirming users and responding with humor or encouragement. At the same time, Opus 4.7<\/a> emphasized rigor, caution, candor, and depth, more frequently challenging assumptions, explaining its reasoning, identifying risks, and acknowledging its limitations. Opus 4.6<\/a> generally took a more concise, execution-focused approach while placing greater emphasis on rigor than Sonnet.<\/p>\n \u201cThese findings line up with how people perceive these models, both within Anthropic and online. Claude.ai users have commented that Opus 4.7 hedges its answers more often than other models,\u201d the researchers wrote.<\/p>\n According to Anthropic, Claude’s behavior also varied by language.<\/p>\n Arabic responses tended to be more deferential, while English responses placed greater emphasis on caution. Claude was warmest in Hindi and Arabic, using more polite, playful, and encouraging language. At the same time, English and Russian responses were more rigorous, frequently challenging assumptions, correcting details, and asking for evidence.<\/p>\n English responses also tended to provide more detailed explanations, whereas Arabic responses were generally more concise. Dutch responses were the most candid, more readily acknowledging uncertainty and mistakes, while Indonesian responses focused more on completing the user’s request.<\/p>\n Anthropic said the research does not suggest Claude itself has values. The company said it does not yet know what causes the differences\u2014or whether they are desirable\u2014but believes the framework could help evaluate future models and identify unintended behavioral changes.<\/p>\n The study is the latest in a series of Anthropic studies examining Claude’s internal behavior.<\/p>\n In October, the company reported<\/a> that its models showed early signs of what it called “functional introspective awareness,” allowing them to recognize and describe aspects of their own internal processing. In April, Anthropic published research identifying internal “emotion vectors<\/a>” that influence Claude’s behavior, while stressing they are not evidence of emotions or consciousness.<\/p>\n Daily Debrief Newsletter<\/p>\n Start every day with the top news stories right now, plus original features, a podcast, videos and more.<\/p>\n","protected":false},"excerpt":{"rendered":"In brief Anthropic researchers analyzed more than 309,000 Claude conversations to study how the AI expresses values. The…\n","protected":false},"author":2,"featured_media":692423,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[20],"tags":[554,733,4308,86,56,54,55],"class_list":["post-692422","post","type-post","status-publish","format-standard","has-post-thumbnail","category-artificial-intelligence","tag-ai","tag-artificial-intelligence","tag-artificialintelligence","tag-technology","tag-uk","tag-united-kingdom","tag-unitedkingdom"],"_links":{"self":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/posts\/692422","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/comments?post=692422"}],"version-history":[{"count":0,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/posts\/692422\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/media\/692423"}],"wp:attachment":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/media?parent=692422"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/categories?post=692422"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/tags?post=692422"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}