Last updated: 14 October 2026
In short
I write the articles myself, in French, for my blog sebcar.net. AI helps me translate and adapt them into English for GateA1, and acts as a proofreader and editing assistant. Nothing is published without my review and approval, and I take responsibility for everything that is online.
Why this page
AI is now part of how I work. I would rather explain clearly how I use it, what it does and what it does not do, than leave any doubt.
In the texts
What I do myself
The choice of topic, the outline, the first draft, the opinions, the anecdotes and the tone are mine. The first draft is written without artificial intelligence, in French, on my blog sebcar.net.
Translation into English
Most GateA1 articles are English versions of my French articles. The translation is done with AI, and it is an adaptation rather than a word-for-word translation: examples, references and wording are adjusted so they make sense to an international reader.
- Every translated article is reviewed and approved by me before publication. I check that the meaning, the facts and my opinions come through unchanged.
- When an article exists in both languages, a link at the bottom of the page leads to the French original.
- Some articles may be written directly for GateA1. In that case, the same rules apply as for my French articles.
What AI helps me with
My main tool is Claude (Anthropic), because it can work directly in the website’s publishing tool. Apart from translation, AI helps with:
- proofreading: spelling, grammar and phrasing;
- fact-checking: it flags claims that are doubtful or out of date. I then check them myself against the sources, because AI can be wrong too;
- updates: when an article ages, it spots outdated information and can suggest a short update;
- editing on the website: links between articles, image alt texts, titles, search engine descriptions and the summaries at the top of articles. It also helped me set up and configure this website.
For these editing tasks, every change is suggested by the AI, then approved by me before it is saved.
What AI does not do
- It does not write original articles in my place, and it does not change their substance or structure. It translates and edits what I wrote.
- It does not publish anything on its own initiative. Articles are published by me, after my review.
- It does not reply to comments or to readers’ messages: I always do that myself.
In the images
I use ChatGPT or Gemini for illustrations. There are two cases.
Generated illustrations
When I have no starting image, the illustration is fully generated from a description (a “prompt”). These images replace royalty-free stock photos: they are unique to the website and fit the topic of each article better. They illustrate a point; they do not document a real event.
Edited photos
When I start from one of my own photos, AI may be used to improve it, fix defects or hide elements that would expose my private life or identity. These edits never change what the photo shows. Screenshots presented as evidence are not altered, except to hide personal information.
Personal data
The AI has no access to readers’ comments or to messages sent through the contact form. I have deliberately blocked this access.
Responsibility and contact
I remain solely responsible for the content published on this website, whether or not it was translated or proofread by AI. If you spot a mistake, or have a question about this page, please write to me through the contact page.