Your intellectual fly is open when you use an LLM to author a post (2025)
Widespread use of large language models to draft posts, especially on platforms like LinkedIn, is raising concerns about authenticity, readability, and intellectual honesty. Many commenters argue that undisclosed AI-generated text amounts to a kind of intellectual catfishing or ghostwriting, eroding trust and depriving both writer and reader of the thinking that happens during real writing. Others see LLMs as legitimate “cognitive prosthetics” or translation tools—especially for non‑native speakers or routine communication—so long as humans still review, endorse, and clearly differentiate between machine‑assisted and machine‑delegated authorship.
LLM Writing Quality & “Slop”
- Many commenters find typical LLM prose verbose, cliché‑ridden, and stylistically homogeneous, especially on LinkedIn and in corporate email.
- Common “tells”: overuse of certain constructions (e.g., “It’s not X, it’s Y”), inflated length, generic enthusiasm, and a flattened, house‑style voice.
- Some argue critics partly want AI output to be bad; others note AI text and images are clearly improving and can already be slick on the surface while remaining shallow.
Authenticity, Disclosure, and Ethics
- Strong view: passing largely LLM‑written text off as your own is misrepresentation or ghostwriting without attribution; akin to plagiarism of a machine.
- Objection to undisclosed use: readers think they’re seeing a person’s thinking, when they’re seeing a model’s template with light prompting.
- Counter‑view: tools are tools; nobody discloses spellcheck, Word grammar, or autocomplete. Authors remain responsible for content regardless of tools used.
- Some propose distinctions like “machine‑assisted” vs “machine‑delegated” authorship and argue for at least high‑level disclosure of how LLMs were used.
Writing as Thinking vs Tool Use
- Many stress that writing forces you to clarify, serialize, and sometimes change your ideas; outsourcing prose means outsourcing part of the thinking.
- Others counter that not all writing is deep thinking (e.g., routine emails, status updates) and AI polish can be pragmatic there.
- Several use LLMs as “cognitive prosthesis” (non‑native speakers, people with health or attention issues), especially for bureaucratic or high‑anxiety communications.
Reader Reactions and Platform Effects
- Numerous people say they now stop reading as soon as they smell LLM style; it signals low effort and disrespect for the reader’s time.
- LinkedIn is repeatedly described as overrun with AI‑generated “thought leadership” and personal‑brand spam, which drives serious readers away.
- Some see this as a “revolt of the reader”: engaged readers prefer flawed but human, context‑rich writing over smooth, generic AI text.
Acceptable and Borderline Uses
- Broadly accepted: grammar and clarity fixes, translation, summarizing paywalled or ephemeral news (with links), internal notes, and purely functional or low‑stakes text.
- Contentious but debated: using LLMs to expand detailed outlines or to rewrite one’s own rough drafts; opinions split on whether this still undermines authenticity.
Future Trajectories and Coding Analogy
- One camp believes that as LLM writing and coding improve, humans will operate at higher abstraction levels (outlines/specs), delegating prose and code generation.
- Another camp argues that current models show fundamental limits in understanding and originality; next‑token prediction drives toward median, de‑personalized output.
- Worry: if undetectable LLM writing and coding become common, identity, trust, and maintainability (of both prose and code) become major long‑term problems.