AI-Generated Images Discourage Me from Reading Your Blog
TL;DR Highlight
AI-generated images on a blog make readers suspect the text is AI-written too
Who Should Read
A developer deciding whether to use AI-generated hero images on a tech blog or docs
Core Mechanics
- AI images make readers doubt the text
- The author says seeing AI-generated images makes him wonder whether the text is also partly AI-generated.
- He finds it especially disappointing on individual blogs rather than corporate ones.
- A bad drawing beats an AI image
- The author says he would rather see a crude Microsoft Paint drawing than an AI image.
- He says his own blog has flaws, but readers know the thoughts are from a real human.
- Personal bloggers should skip AI images
- The author's conclusion is that personal blogs should avoid AI-generated images.
- The argument rests on personal feeling, with no numbers or experiments in the article.
Community Reactions
- Images that explain content are fine, one commenter argued
- A commenter said rejecting an image that helps explain the content just because of how it was made is superficial.
- The problem is images used as a false signal of effort.
- AI images lack taste, one commenter said
- A commenter said AI images rarely show taste, which makes them off-putting.
- Stock images carry the artist's fingerprints and the choice itself reveals something about the author.
- A commenter flagged an AI image to marketing and it got replaced
- A commenter criticized an AI image on their company's dashboard, and the team swapped it for an abstract image.
- Salespeople later privately thanked them.
- A commenter's fan-fiction images went unnoticed
- A commenter said their wife uses AI images for fan-fiction scenes and most readers don't notice.
- Using ElevenLabs for narration made AI use obvious and cost her subscribers.
How to Apply
- Audit the purpose of each blog image
- Today, write one line per post on whether the image explains something or just fills space.
- Remove filler images or replace them with diagrams you make yourself.
- Replace with your own sketches or screenshots
- Rough hand-drawn diagrams on a tablet or pen are fine.
- Use real screen captures or code output so the image ties directly to the text.
- Disclose it if you use AI images
- Put the tool used in the image caption.
- State whether the text was written by you to reduce reader suspicion.
Terminology
Related Papers
AI is a tool and using it well takes skill
A rebuttal to Hickey: LLMs are tools, and skill in directing and checking them decides the outcome
Reflective Prompt Optimization of a Deep-Research Agent: A Topic-Disjoint, Multi-Seed Protocol for Measuring Transferable Gains and Prompt-Level Overfitting
Reflective prompt optimization raised held-out citation reward from 0.702 to 0.781 on a deep-research agent.
GPT‑6 and Intelligent UI for everyone
GPT-6 reportedly adds generated UI to chat; with no article text, only comment reactions are available.
Base Models Can Reason By Taking a Cue From Training Data
Prefilling just two opening tokens lifted Olmo-3-7B's MATH-500 accuracy from 42% to 78%.
Labels Override Definitions in Jev-Style Typed Decision Models
Renaming options to A/B raised routing accuracy by +0.1511 over meaningful labels.
How Much Prompt Is Enough? A Blackbox Minimization of Few-Shots in LLMs
Delta debugging trimmed few-shot prompts by 65.3% on average with unchanged propositional output