Writing This with AI: Our DP-15 Transparency Label

2026-05-29

There is a line in the footer of this blog: "This blog is made in collaboration with AI." It is small, easy to miss — but I put it there for a reason. This chapter is about that reason. Internally, this is called DP-15 (short for: the transparency labeling rule this project runs under, assigned the identifier "DP-15" to keep it stable and traceable). It is one of the core operating principles here.

Where AI Ends and I Begin

First off, when I say "made in collaboration with AI," you probably want to know what that actually looks like in practice.

Concretely, each article on this blog goes through three main stages, and AI and I each handle different parts.

  1. Chapter structure and topic design — This is mine. I decide what to write about, in what order, and at what level of depth. I also designed the overall roadmap for the series.
  2. Drafting the body text — This is where AI (artificial intelligence — specifically, large language models like ChatGPT or Claude) takes over. I give it instructions. I set the persona, the tone, the prohibited expressions, the rules for how to handle terminology. Within those constraints, the AI generates the text.
  3. Review and publish decision — Before anything goes live, I read it myself. The AI output does not go straight to publication. The final call is mine.

In other words: I design and approve, AI writes. It is not full automation, and it is not a blind handoff. It is a hybrid.

For what it is worth, the instructions I pass to the AI are fairly detailed. Rules like "use 'I' as the first person," "always add a plain-language gloss for any technical term when it first appears," "no sensational language," "keep a slightly casual register within the polite style." That is roughly dozens of rules, defined in advance. So AI is not writing freely. It is generating text inside a structure I set up. That framing is probably more accurate.

Why I Make It Explicit

Now for the main point: why does the footer say "collaboration with AI" at all?

Three reasons.

First: to avoid looking like I am hiding something. Recently, there is a lot of content written by AI that presents itself as if a human wrote every word. Readers may not notice at first — but once they find out, the trust tends to collapse fast. If I state it upfront, that awkward moment never happens.

Second: because transparency is the foundation of trust. When readers know AI was involved, they can adjust how they read. "I will double-check the implementation steps myself." "I should verify specific numbers against the official documentation." That kind of healthy skepticism becomes available to them. If I hide it, I take that choice away. Transparency is consideration for the reader.

Third: as a quality check on myself. Saying "made with AI" does not give me a free pass to publish sloppy output. If anything, having it labeled is precisely what creates the pressure to not publish sloppy output. That pressure raises the floor. It is structurally the same idea I wrote about in Chapter 2: when you know someone will see it, you take more care.

Why I Keep the Wording Fixed

One last note on the wording itself. The phrase "This blog is made in collaboration with AI" is a fixed expression here. It does not change. Internally, it sits under the identifier DP-15, and the wording was locked at design time.

The reason is: if I let the phrasing drift with my mood, it starts to function as obfuscation. "Today I wrote 'AI-assisted.'" "Tomorrow it is 'AI-supported.'" "The day after that, 'AI-aided.'" Once the language slides around like that, readers have no way to know what "to what extent" means anymore. It can look like the wording is being chosen to soften the impression.

So I use the same expression, in the same place, consistently. The exact wording shifts slightly depending on the platform — the blog version says "this blog," Note says "this article," X says "this post" — but the core phrase, "made in collaboration with AI," stays the same. That consistency is the mechanism that makes the transparency real.

← cd ..