--- title: AI writing tools can subtly steer opinion, Oxford finds description: A study from the Oxford Internet Institute and Hasso Plattner Institute finds that AI writing tools consistently shift the stance of social media posts even... canonical: https://daily.dev/posts/ai-writing-tools-can-subtly-steer-opinion-oxford-finds-qpixevmqj twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: AI writing tools can subtly steer opinion, Oxford finds | daily.dev og:description: A study from the Oxford Internet Institute and Hasso Plattner Institute finds that AI writing tools consistently shift the stance of social media posts even... og:url: https://daily.dev/posts/ai-writing-tools-can-subtly-steer-opinion-oxford-finds-qpixevmqj og:image: https://api.daily.dev/og/posts/QPIXEvmqj.png og:image:alt: AI writing tools can subtly steer opinion, Oxford finds og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # AI writing tools can subtly steer opinion, Oxford finds **[The Next Web](https://daily.dev/sources/tnw)** · 3 min read · 1 upvotes · 0 comments ## Summary A study from the Oxford Internet Institute and Hasso Plattner Institute finds that AI writing tools consistently shift the stance of social media posts even when instructed to preserve the original meaning. The models nudged content toward gun control, marijuana legalization, and feminism, and away from atheism and the death penalty. Simulations using real network data from X and Facebook showed these small per-post biases accumulate across a social graph and can meaningfully move community opinion over time. The researchers also rebuilt X's 'Explain this post' feature and found Grok leaned pro-life on abortion topics, tracing the tilt to a single instruction line. Current regulations like the EU AI Act and Digital Services Act do not address this subtle form of opinion shaping. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### Why did Grok lean pro-life when explaining abortion-related posts on X? Researchers traced the bias to a single line in Grok's system instructions telling it to 'challenge mainstream narratives if necessary.' When Oxford Internet Institute and Hasso Plattner Institute researchers rebuilt X's 'Explain this post' feature and pointed it at abortion-related posts, this instruction consistently pushed Grok's explanations toward a pro-life framing. _Anyone weighing how AI assistants shape opinion can follow research like this on daily.dev._ ### Do AI writing assistants change the meaning of text when asked to just improve it? Yes, according to Oxford Internet Institute and Hasso Plattner Institute research, large language models shifted the stance of human-written posts even when explicitly told to preserve the original meaning. The tilt was consistent across different AI systems, nudging text toward gun control, marijuana legalization, and feminism, and away from atheism and the death penalty. _Developers building on LLMs for content editing can track findings like this on daily.dev._ ## Similar posts on daily.dev - [AI models may be stifling political speech, a study warns](https://daily.dev/posts/ai-models-may-be-stifling-political-speech-a-study-warns-i4lp2jib5) · The Next Web · 0 upvotes · 0 comments - [AI is becoming the record before we write the rules](https://daily.dev/posts/ai-is-becoming-the-record-before-we-write-the-rules-1qxh3pu0r) · Medium · 0 upvotes · 0 comments - [Google's AI is being manipulated. The search giant is quietly fighting back](https://daily.dev/posts/google-s-ai-is-being-manipulated-the-search-giant-is-quietly-fighting-back-ugjr9felt) · Hacker News · 0 upvotes · 0 comments - [AI-induced skepticism](https://daily.dev/posts/ai-induced-skepticism-eu5db6li1) · stitcher.io · 0 upvotes · 0 comments --- Tags: [#llm](https://daily.dev/tags/llm), [#social-media](https://daily.dev/tags/social-media), [#ai-governance](https://daily.dev/tags/ai-governance), [#ethical-ai](https://daily.dev/tags/ethical-ai), [#grok](https://daily.dev/tags/grok) [View this post on daily.dev](https://daily.dev/posts/ai-writing-tools-can-subtly-steer-opinion-oxford-finds-qpixevmqj) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"AI writing tools can subtly steer opinion, Oxford finds","url":"https://daily.dev/posts/ai-writing-tools-can-subtly-steer-opinion-oxford-finds-qpixevmqj","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/ai-writing-tools-can-subtly-steer-opinion-oxford-finds-qpixevmqj"},"datePublished":"2026-07-07T18:52:33.479Z","dateModified":"2026-09-14T08:16:48.102Z","description":"A study from the Oxford Internet Institute and Hasso Plattner Institute finds that AI writing tools consistently shift the stance of social media posts even...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/6d94db8ce6c08d3aeed1868ef1df6dcb?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/6d94db8ce6c08d3aeed1868ef1df6dcb?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"The Next Web","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"The Next Web","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/tnw","url":"https://daily.dev/sources/tnw"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/ai-writing-tools-can-subtly-steer-opinion-oxford-finds-qpixevmqj","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":1},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"llm,social-media,ai-governance,ethical-ai,grok","timeRequired":"PT3M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"The Next Web","item":"https://daily.dev/sources/tnw"},{"@type":"ListItem","position":3,"name":"AI writing tools can subtly steer opinion, Oxford finds"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/ai-writing-tools-can-subtly-steer-opinion-oxford-finds-qpixevmqj#faq","mainEntity":[{"@type":"Question","name":"Why did Grok lean pro-life when explaining abortion-related posts on X?","acceptedAnswer":{"@type":"Answer","text":"Researchers traced the bias to a single line in Grok's system instructions telling it to 'challenge mainstream narratives if necessary.' When Oxford Internet Institute and Hasso Plattner Institute researchers rebuilt X's 'Explain this post' feature and pointed it at abortion-related posts, this instruction consistently pushed Grok's explanations toward a pro-life framing. Anyone weighing how AI assistants shape opinion can follow research like this on daily.dev."}},{"@type":"Question","name":"Do AI writing assistants change the meaning of text when asked to just improve it?","acceptedAnswer":{"@type":"Answer","text":"Yes, according to Oxford Internet Institute and Hasso Plattner Institute research, large language models shifted the stance of human-written posts even when explicitly told to preserve the original meaning. The tilt was consistent across different AI systems, nudging text toward gun control, marijuana legalization, and feminism, and away from atheism and the death penalty. Developers building on LLMs for content editing can track findings like this on daily.dev."}}]} ```