· Bullet points for everything · single-tell paste-in
Stop ChatGPT from using bullet points.
LLMs default to bullet lists because they're easy to generate and easy to skim. They make replies feel like slides. Paste the prompt below and the model writes in paragraphs, reserving bullets for genuinely parallel items.
· the paste-in
Apply this rule to every reply you write for me. It removes a specific pattern that flags content as AI-written. Rule: Default to prose paragraphs. Only use a bulleted list when the items are truly parallel and don't connect via cause/effect or sequencing. If a list has only two or three items, see if the prose form reads better. Self-review: Count bulleted lists in the output. If more than one third of the content is bulleted, convert some lists back to prose. Before sending any reply, run through the self-review. If the rule was violated, rewrite the affected sentence from scratch rather than character-swapping.
What this tell is
AI over-bullets when the underlying ideas connect via cause-and-effect, sequencing, or argumentation but the model breaks them into discrete bullets anyway. Symptoms: three-bullet lists where two are real and one is padding, bullets for ideas that have a logical flow, bullets repeated across every section of a longer reply.
Why removing it matters
Bullet-heavy replies read as AI-shaped because human writers default to prose for connected ideas and reserve bullets for genuinely parallel lists (ingredients, options to compare, distinct steps). When everything is bullets, nothing has emphasis and the reader has to do the connecting work themselves.
Per-model notes
ChatGPT GPT-5.5 is the most over-bulleted by default; Markdown-heavy output is its house style. Claude 4.7 over-bullets less but still defaults to lists for technical answers. Gemini varies by query.
· want the full set?
This page covers bullet points for everything only. To kill 20 other 2026 AI tells (em dashes, the word "comprehensive", AI vocabulary cluster, templated transitions, sycophancy, hallucination, rule-of-three lists, over-bolding, and more) in one paste-in prompt, plus a Claude Code rules block and a Node.js lint script: use the AI Tell Killer main tool.
FAQ
Will this make code answers harder to read?
No. Code blocks and step-by-step technical instructions (where each step is a discrete action) are still appropriate uses of structure. The rule targets unnecessary bullets in prose: bulleting two or three things that would read better as a sentence. Models keep code blocks, numbered installation steps, and genuine option-comparison lists.
What if I actually want a bulleted list?
Ask for one. 'Give me a bulleted list of...' overrides the rule for that turn. The rule changes the default behavior, not your ability to request structure when you need it.
Related: the full AI Tell Killer (all 21 tells, three output formats), the AI Output Linter (paste AI text and scan for tells in real time), and the em dashes guide (deep-dive on the single most-checked tell).