How LLMs Distort Our Written Language

大型语言模型如何扭曲我们的书面语言How LLMs Distort Our Written Language本文介绍了一项由加州大学伯克利分校、加州大学圣地亚哥分校、华盛顿大学、宰图纳学院以及谷歌DeepMind的研究人员于2026年3月联合发表的研究。该研究指出，全球每周有超过十亿人使用大型语言模型辅助写作，但这些模型不仅改变了人类写作的语气和语调，还持续地改变了文本的预期含义。研究表明，人工智能的广泛使用可能会引导人类的写作和决策偏离其最初的意图，从而对文化和科学机构产生深远的影响。This article introduces a study jointly published in March 2026 by researchers from UC Berkeley, UC San Diego, the University of Washington, Zaytuna College, and Google DeepMind. The research points out that over a billion people globally use large language models weekly to assist with writing, but these models not only alter the voice and tone of human writing, but also consistently alter the intended meaning. The study indicates that the widespread use of artificial intelligence may subtly steer human writing and decision-making away from original intentions, thereby having profound impacts on cultural and scientific institutions.感谢您阅读本期内容。我们全面介绍了大型语言模型对书面语言影响的最新研究成果。如果您对数据分析、人工智能以及更多前沿科技内容感兴趣，诚邀您关注Learn By Doing With Steven 数能生智、steven data talk、steven数据漫谈等系列节目。相关内容已同步在小红书、微信公众号、YouTube、Spotify等平台更新。Steven Data Talk播客也可在YouTube Music和Spotify收听。欢迎访问节目描述区域的Linktree链接，获取我们所有社交媒体平台的完整列表。Thank you for reading this issue. We have comprehensively introduced the latest research findings on the impact of large language models on written language. If you are interested in data analysis, artificial intelligence, and more cutting-edge technology content, we sincerely invite you to follow Learn By Doing With Steven, steven data talk, steven数据漫谈, and other related programs. Relevant content is updated across platforms including Xiaohongshu, WeChat Public Account, YouTube, and Spotify. The Steven Data Talk podcast is also available on YouTube Music and Spotify. Please visit the Linktree link in the show notes area to access the complete list of our social media platforms.https://arxiv.org/pdf/2603.18161https://linktr.ee/learnbydoingwithsteven#人工智能 #大型语言模型 #科学研究 #写作助手 #数据漫谈 #LearnByDoingWithSteven #StevenDataTalk