Why Every Post on My Blog Has an AI Badge

Scroll to the bottom of any post on my blog and you'll find a small badge. It tells you how much AI went into what you just read.

Go back far enough and they all say No AI. Everything I wrote between 2018 and 2021, the climate posts, the campervan trip, my year reviews, was typed by me.

Everything since I restarted the blog in 2025 says AI Assisted. Every single one.

I didn't plan to draw a line through my own archive. But it's there, and I'd rather point at it than pretend it isn't.

How I ended up here

When I revived the blog, I co-wrote the first post with ChatGPT. I fed it my old posts, let it interview me, answered out loud with a dictation tool, then edited what came back. It sounded like me. It was me, mostly.

Straight after, I wrote Did I Write That Blog Post? because I genuinely wasn't sure. My answer was “kind of.” I still think that's the honest one.

Everyone's writing with AI now, and we all know it. What we don't know is how much. Is that post a person with a spell checker, or a prompt and a publish button? From the outside, you can't tell.

Why a badge and not a disclaimer

You could put a line at the top of every post. I didn't want to. “This post was written with help from AI” reads like an apology, a warning label stuck on something I'm proud of.

A badge does the same job without the drama. It's small, it sits in the same place every time, and you learn what it means after seeing it once. It reads more like a nutrition label than a confession.

So I made three, and put them on AI Honesty Badge for anyone to use:

• No AI: made by a person, start to finish
• AI Assisted: I led the work and AI helped along the way
• AI Generated: AI made it from a brief, and I published roughly what came back

Where I draw the line

The hard part is being honest about the middle.

My rule for AI Assisted is simple. The idea is mine, the stories and opinions are mine, and I've edited it until it sounds like me. If all of that's true, it's AI Assisted, even when AI wrote the first draft.

That last bit might surprise people, but it's how most of my recent posts were made.

Spell check doesn't count. Handing ChatGPT a topic and publishing whatever comes back definitely does. That's AI Generated, and I'd say so.

You'll draw your lines in slightly different places, and that's fine. The badge is a promise you make to your readers.

Why admitting it builds trust

You'd think a badge like this would put readers off. I think it does the opposite.

When someone goes out of their way to add an AI Assisted badge to a post, I trust them more. They didn't have to. They told me anyway, and I appreciate that.

The alternative is guessing. I've noticed it in myself: the more AI stuff turns up in my feeds, the less I trust any of it. A badge won't fix that on its own. But it's one small, honest signal in a feed that's short on them.

Three badges, not two

Most people will reach for No AI, and that's needed. There's real value in being able to say a person made this, start to finish.

I don't expect many people to use AI Generated. Some will be embarrassed to. That's exactly why it should exist. If the honest answer is “AI made it”, there should be a way to say it that doesn't feel like a confession.

AI Assisted sits in the middle, which is where I think most of us actually are.

It runs on trust

There's no verification. Anyone can stick a No AI badge on something ChatGPT wrote.

I know. It works like an honesty box at a farm gate. Nobody's watching, most people pay anyway, and the few who don't are mostly cheating themselves.

What's surprised me is how differently people use them. Baystation12, a long-running Space Station 13 game server, puts the No AI badge in its README under one blunt line: “Generative AI is not welcome here.” Jake Cozart labels each of his open-source projects separately, some AI Assisted, some AI Generated. Mark Hall adds one to posts where AI made the images or checked his logic.

They take very different positions on AI and use the same three badges to say so, which is what I hoped for.

Adding one to your own work

Pick the badge that's honest, copy the snippet, and paste it wherever you want it to show. There's HTML for websites and Markdown for GitHub READMEs. It's free for anything, including commercial work, and there's nothing to install.

If you're not sure which one fits, the No AI, AI Assisted and AI Generated pages each explain where the lines sit. And if you use one, tell me. I'll add you to the Honesty Wall.

Or let your AI do it. If you write with Claude Code, I built a skill that adds the badge for you. It knows how much it helped, so it picks the badge from what actually happened in the session. It follows however your blog already does things, and it won't label its own work No AI. If you use Cursor, Codex or something else, there's a line you can add to your AGENTS.md that does the same job. Both are on the AI tools page.