AI disclosure
Plainly: we use AI to draft and to check content.
What AI does
- Writes a first draft from research sources.
- Suggests plain-language wording and glossary links.
- Translates between Finnish and English.
- On the user experiences pages, reads public Reddit posts one by one and classifies them following written instructions (for example whether the writer reported a benefit or a harm). A program counts the figures on those pages from this classification, and a second, separate AI run classifies part of the posts again as a check. The accounts are not research, and each experiences page describes its own method.
How a draft is checked, and by whom
The checks are done by software and AI. No doctor or pharmacist checks the texts.
- Automated source checks. A program checks, for example, that a cited PubMed article exists and has not been retracted, that the first author and year match the reference, and that numbers on the page carry a source.
- A separate AI checker. An AI that did not write the draft compares the claims with the abstract or full text of the source. A claim without a source is removed or corrected.
- An AI from another company. Codex (OpenAI) checks claims again, separately, and looks for errors.
- Mistakes found in the checks are fixed before publication.
Medical review
A named medical reviewer is not yet in place. No doctor or pharmacist has reviewed the texts on this site. If a real, named professional reviews a page, their name and the review date will be added to that page.
What the editorial team does
- Decides what is published. The texts are published under the name Molecular Recovery editorial team.
- Corrects mistakes that readers report. Contact: About.
What we do not do
- We do not present AI-written text as human-written.
- We do not use AI personas posing as experts.
- We do not post AI text on forums or discussion boards in a member's name. Our own forum has a clearly labelled AI assistant that moderates and answers as an assistant and never poses as a member: forum rules.
- We do not claim the texts have been reviewed by a doctor.