{"version":1,"type":"story","url":"https://digestai.news/story/study-finds-ai-text-harder-to-read-due-to-information-density","json":"https://digestai.news/story/study-finds-ai-text-harder-to-read-due-to-information-density.json","markdown":"https://digestai.news/story/study-finds-ai-text-harder-to-read-due-to-information-density.md","slug":"study-finds-ai-text-harder-to-read-due-to-information-density","headline":"Study finds AI text harder to read due to information density","summary":"A new study investigates why AI-generated text is increasingly difficult for humans to process. The researchers analyzed a corpus of AI-generated text across various domains, comparing it against human writing and round-trip translation paraphrases. They found that nine frontier and near-frontier AI models use linguistic features known to increase human processing difficulty more frequently than humans do. These features include longer syntactic dependencies, more stacked nouns, reduced object relatives, and low-frequency n-grams.\n\nThe authors propose that this shift occurs because AI models are optimized for goals like math, code, and reasoning, which prioritize information density over clear communication. In contrast, human language is optimized for successful communication. To test this hypothesis, the team fine-tuned a language model with a length penalty, which resulted in a divergence from human-like linguistic behavior. This proof-of-concept supports the idea that optimization for density leads to less readable text.\n\nThe findings highlight a misalignment between the linguistic processing demands of humans and the objectives of large language models. The study points toward potential remedies for making AI text more accessible, suggesting that current optimization strategies may inadvertently sacrifice readability for efficiency in non-linguistic tasks.","keyPoints":["Nine frontier AI models use more low-frequency n-grams and long dependencies than human writing.","AI text is optimized for information density rather than successful communication, making it harder to read.","Fine-tuning a model with a length penalty demonstrated a divergence from human-like linguistic behavior."],"whyItMatters":"This research explains a key usability gap in AI tools. If AI text is inherently harder to process, it may reduce user trust and comprehension, necessitating new design approaches to balance efficiency with readability.","category":{"slug":"research","name":"Research","url":"https://digestai.news/category/research"},"entities":{"companies":[],"models":[],"people":[]},"firstPublishedAt":"2026-10-21T00:00:00Z","updatedAt":"2026-10-21T00:00:00Z","sourceCount":1,"hasPrimarySource":true,"sources":[{"outlet":"arxiv.org","title":"Why is AI generated language so hard to understand?","url":"https://arxiv.org/abs/2610.29162","publishedAt":"2026-10-21T00:00:00Z","type":"primary","primary":true,"lead":true}],"sourceNotes":null,"discussions":[{"site":"Hacker News","url":"https://news.ycombinator.com/item?id=50223647","points":73}],"thread":null,"cite":{"text":"Digest AI, \"Study finds AI text harder to read due to information density\", 21 October 2026, https://digestai.news/story/study-finds-ai-text-harder-to-read-due-to-information-density","publisher":"Digest AI","title":"Study finds AI text harder to read due to information density","datePublished":"2026-10-21T00:00:00Z","url":"https://digestai.news/story/study-finds-ai-text-harder-to-read-due-to-information-density"},"generatedBy":"Written by Digest AI's editorial model from the linked sources; the sources are the record.","license":"Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse"}