Last updated: August 2, 2026
Desire Index publishes educational content about desire in long-term relationships. That subject sits close to people's health, their marriages, and their sense of self, so the bar for accuracy is high. This page explains how our content is made, who reviews it, how citations are checked, and what happens when we get something wrong.
Our drafts are AI-assisted. We disclose this plainly because it is true and because you deserve to know how the words in front of you were made. A language model produces the first draft of most articles, working from a structured brief and a library of vetted research claims. The model is a drafting tool. Every page is reviewed by a named human before it is published, and every citation is verified against the source. No page goes live on the strength of the model alone.
Citations on this site must resolve. Journal articles are linked through their DOI, and the link leads to the published work itself. Claims enter our content only through a vetted claims library: each entry records the finding, the study it comes from, and the exact scope of what that study supports. A sentence that stretches a claim beyond what the source says is flagged during review and either softened or cut.
Every page passes a structured human review before publish. Editorial responsibility rests with Seaward Labs, LLC, the company that develops Desire Index. The review follows a 20-minute protocol: reading the title, intro, and headings aloud to check the register, opening randomly selected citations and confirming each sentence fairly represents its source, scanning for tone and for proper signposting on any crisis-adjacent material, checking that the reader's path through the page makes sense, and weighing any warnings raised by our automated checks. A page that fails review is revised or killed.
We plan to engage a credentialed clinical reviewer, such as an AASECT-certified sex therapist, a licensed clinical psychologist, or a licensed marriage and family therapist with visible work in sexual desire, before this site grows past roughly 30 articles. When that reviewer signs on, reviewed pages will name them and their credentials, and a review log will record what they reviewed and when. We will never attach a fabricated reviewer to a page.
If you find an error, write to us at info@seawardlabs.io. We investigate every report against the original source and correct substantive errors promptly. Substantive edits bump the visible updated date on the page, so the date you see always reflects the last time the content meaningfully changed.
Evergreen articles are reviewed at least once a year. Pages built on statistics are reviewed twice a year, because the datasets behind them update on their own schedules. If a review changes the substance of a page, the visible updated date moves. If it confirms the page is still accurate, the date stays put.
The research framework behind the assessment is documented in our methodology. Who we are and why this exists is on the about page.