How we use artificial intelligence

Last updated: 17 October 2026

To keep the site safe and free, an artificial intelligence system checks some content before it's published. Here we explain what it checks, what it never does and how you can ask for a person to decide.

What it checks

  • the title and text of ads;
  • members' stories (with their own rules: the text can be explicit, but the characters must be consenting adults);
  • the profile description;
  • profile and album photos.

What it looks for

  • offers or requests for payment: money, gifts, “tributes”, top-ups, selling photos or videos;
  • references to minors and photos in which minors might appear;
  • explicit nudity where it isn't allowed (profile photos and albums visible to everyone);
  • violence, illegal content, contact details written in the text or in photos;
  • in stories: minors, even merely implied, relatives, violence or non-consensual acts, real people (“forced” domination play only passes if the text shows it was agreed, with a safeword).

What happens after the check

  • Approved: the content is published.
  • Uncertain: a person examines it before publication.
  • Rejected: it isn't published and we explain why.

Two cases have consequences for your account: a photo with a possible minor immediately suspends the account pending a check by a person; three rejections in 30 days for requests for payment suspend the account. Every decision comes with reasons and you can appeal it within 6 months from the Decisions page: it is always reviewed by a person.

What it never does

  • it doesn't read messages or look at chat photos (a reported message is seen only by a person, the moderator);
  • it doesn't decide which ads you see or in what order, and doesn't profile you;
  • it doesn't use your content to train artificial intelligence models;
  • it doesn't make final decisions without you being able to ask for a person to step in.

Who provides the system

We use Google's Gemini model through the paid Gemini API service. Google doesn't use the text and photos we send to improve its products; it keeps them for 55 days solely to detect abuse of its policies and, in those cases, they may be examined by its authorised staff. Processing may take place in the United States, with GDPR safeguards (see the privacy notice).

We keep the outcome of each check for 12 months, so we can respond to appeals and improve the rules.

References

Regulation (EU) 2016/679 (GDPR), arts. 13, 14 and 22; Regulation (EU) 2022/2065 (DSA), arts. 14, 17 and 20; Regulation (EU) 2024/1689 on artificial intelligence. The content rules are in the Rules.