This page says exactly where AI tools are used in producing project-diada.org, and where they are not. We would rather be plainly honest about it than leave readers to guess.
Yes, we use AI tools
project-diada.org uses AI language models as part of its production process. Pretending otherwise would be dishonest, and readers of a site about human behaviour deserve better than that.
What AI does here is research assistance and drafting. What it does not do is decide what is true.
Where AI is used
- Scanning research sources. Filtering incoming journal and research feeds to surface studies within our subject areas.
- Summarising papers. Producing a working summary of a study’s method, sample and findings for the editor to check against the paper.
- Drafting. Producing a first draft of an article from verified source material, following our written style and evidence rules.
- Checking our own drafts. Flagging overstated language, missing caveats and causal verbs attached to correlational findings.
Where AI is not used
- It is not the source of any fact. Every factual claim traces to a paper or an official dataset, located and checked independently. Nothing is published because a model asserted it.
- It does not generate quotations. Every quote comes from a real, named person in a locatable source. No model output is ever presented as something a person said.
- It does not invent people or examples. No fictional patients, no composite readers, no illustrative anecdotes about people who do not exist.
- It does not decide what runs. Commissioning, rejection and publication decisions are human.
- It does not publish. Nothing reaches the site without a human editor reading it first.
- It does not generate our images. Feature images are real photographs, licensed and credited.
Human review is not optional
Every article is read, edited and approved by a human before publication. That pass checks the claims against the sources, the language against the study design, and the piece against our safety rules. A draft that has not been through it does not get published, and Ethan Parker, Editor is accountable for what does.
An AI draft is a starting point. The judgement about what matters, what to cut, and whether a finding is being oversold is the part that makes an article worth reading, and it is not automated here.
Why we disclose this
Search engines judge content on quality rather than on whether a machine was involved in producing it, so disclosure costs us nothing in ranking terms. We do it because a site asking readers to trust its handling of evidence cannot be evasive about its own methods.
What this means for you
Judge our articles the way you would judge any other publication’s: check whether the sources are linked, whether the caveats are present, whether the claims match the studies. Those links are there so you can. If our work does not hold up to that, we would like to hear about it at [email protected].
Changes to this policy
If our use of these tools changes materially, this page is updated and dated. Last updated August 22, 2026.