Sophie Alpert argues that AI-generated text currently cannot reliably preserve an author’s intended meaning without careful human review and editing. She emphasizes that writing is an essential thinking process, and authors must fully own and understand every part of their writing to communicate clearly and respect readers' time. While AI can assist in brainstorming or proofreading, authors should avoid relying on it for unedited content and instead use it as a tool to support, not replace, their own detailed thought and expression.

https://sophiebits.com/2026/06/25/there-are-no-lossless-transformations-of-natural-language-text

The article argues that large language models (LLMs) have contributed to the decline of traditional paragraph writing online, promoting fragmented, “spaghetti writing” styles characterized by excessive line breaks and bullet points that disrupt cohesive thought. This trend, once prevalent only in social media, is spreading to other online spaces where thoughtful, contiguous paragraphs were once the norm, diminishing the quality of written communication. The author encourages writers to relearn paragraph structure to foster clearer, more meaningful writing and support a more literate digital culture.

https://arhan.sh/blog/bring-back-the-paragraph/

The author argues that using AI to write substantive texts—such as essays, reports, or novels—is generally unwise because the writing process is integral to thinking, AI-generated writing tends to be vague and subtly incorrect in ways that are hard to detect, and presenting AI-written text without disclosure is misleading. While AI can be helpful for tasks like brainstorming, editing, or data analysis, relying on it to produce core content can compromise clarity, accuracy, and intellectual rigor, and should only be done with clear labeling to maintain ethical transparency.

https://www.erichgrunewald.com/posts/why-i-think-you-should-almost-never-use-ai-to-write-anything-substantive/