AI-written LinkedIn posts risk eroding readers’ trust, blog argues
The Observation Deck post says excessive reliance on large language models can make LinkedIn writing appear artificial and weaken confidence in the author’s voice.
Heavy use of large language models to write LinkedIn posts can make content recognisably artificial and undermine its authenticity, according to a post published by The Observation Deck.
The author identifies heavy emoji use, single-sentence paragraphs, formulaic constructions and frequent em dashes as stylistic features associated with AI-generated LinkedIn writing. The post cautions that these cues are not definitive proof of LLM authorship, but argues they may be increasingly familiar to readers.
The concern, the author writes, extends beyond awkward style. When a post no longer sounds like its author, readers may question whether its underlying ideas are genuine. The author says this uncertainty can lead them, and others, to stop reading.
The post describes LinkedIn as an important social network for the author, despite calling it comparatively clumsy and conventional. It was originally published as a LinkedIn post on 11 November 2025 and later appeared on The Observation Deck on 5 December under the headline “Your intellectual fly is open”.
The author does not reject LLMs outright, recommending them instead for brainstorming, comprehending text and editing. The final writing, the post argues, should remain in the writer’s own voice.
