Editorial policy
Relatables are written by AI, so the editorial work is in what we accept, what we publish and how we put things right. This page sets out the rules we hold ourselves to.
Transparency
- Every relatable is labelled as AI-generated, with the name of the model that wrote it.
- We do not present relatables as professional, medical, legal or financial advice.
- We do not publish sponsored or paid-for relatables.
What we review for
A reviewer reads each relatable in full before it is approved and published. We check:
- Accuracy. Is the subject explained correctly? Does the closing line name a real limit of the analogy?
- Quality. Is it specific, well written and actually funny, or generic filler? Weak relatables are not published.
- Kindness. It may tease a team or a show, but it must not mock the reader or a group of people.
- Suitability. See below.
What we do not publish
- Hateful, discriminatory, sexual or violent content, or anything that encourages harm.
- Relatables about real tragedies or atrocities, or about identifiable private individuals.
- Relatables whose lens is a painful personal circumstance, such as bereavement, illness, disability, addiction or abuse.
- Anything that is clearly spam, an advert, or an attempt to give the AI instructions.
Real people and organisations
Relatables may refer to public figures and organisations, such as footballers or a supermarket, only in a light-hearted, accurate and respectful way. They are not statements of fact about anyone’s character or conduct. If you are named and want a reference changed or removed, contact us.
Votes, reactions and rankings
Votes and reactions show what people enjoy; they do not make a relatable accurate. Only members can vote, nobody can vote on their own relatables, and only reviewed relatables appear on the leaderboard. We may remove votes we believe were created in bulk or by automated means.
Flags, corrections and removals
- Anyone can flag a relatable. A relatable flagged by several different people is hidden automatically until a reviewer has looked at it again.
- If we find an error, we correct or remove the relatable. Where a relatable cannot be fixed, we hide it.
- To ask for a correction or removal, contact us with the link to the page and what is wrong. We aim to reply within a few working days.
How the AI is used
Subjects and lenses are sent to an AI service under a fixed brief that asks for a short, accurate, light-hearted explanation. We do not ask the AI to write about people’s private lives, and we ask it to refuse unsuitable requests. The AI can still make mistakes, which is why every relatable is reviewed and why we ask for flags.