AI transparency
Every project and piece of writing on this site carries two badges: one for the code, one for the text. They say how much of the work an AI model did. The two are rated separately, because they are rarely the same.
The three levels
- manual
- No AI, or only at the level of search and question-and-answer.
- assisted
- An AI suggested things or produced parts of the work. The design, the decisions and the review are mine.
- generated
- An AI produced most of it. I directed the work and verified the result.
Two separate ratings
- Code
- Code covers everything that runs: source files, scripts, configuration.
- Writing
- Writing covers the prose on the page: the explanation, the analysis, the commentary.
What the badges look like
- Code manual
- Writing manual
- Code assisted
- Writing manual
- Code generated
- Writing manual
Translations
This site is written in English and Turkish. A translation that an AI drafted and I have not reviewed yet is marked separately from one I have gone over.
- Translation AI draft
Why say it at all
A reader deciding whether to trust a piece of work deserves to know how it was made. Hiding AI involvement would make every page a little less worth reading.