How does your work end up cited on Wikipedia? Not by you.

Search “how to get cited on Wikipedia” and almost every result answers a different question: how to cite Wikipedia in your own paper, or how to edit an article. The thing you actually want, getting your research used as a source by someone else, barely comes up. So here is the guide for that.

TL;DR: A Wikipedia citation is now an AI-visibility asset, because Wikipedia is one of the most heavily weighted sources in the data behind large language models. You can’t add your own work to get one: self-citation is a conflict of interest that editors watch for and revert. The honest path is to make your work the kind of source an editor reaches for, and to improve an article in your field without writing yourself into it.

Why does a Wikipedia citation matter more than its traffic?

Because the readers include the machines. Wikipedia shows up over and over in the corpora used to train large language models. It is a substantive part of Common Crawl and of The Pile, and it made up more than half of the data behind the original BERT model. In the GPT-3 training mix (Brown et al., 2020, Table 2.2), Wikipedia carried the highest sampling weight of any single source: tiny in raw size next to the web crawl, yet the model was pointed at it roughly 3.4 times, more than at anything else.

Read that again from your side of the desk. A sentence that cites your paper on a Wikipedia article isn’t just read by the people who land on that page. It is read, and reread, by the systems that increasingly answer questions for everyone else. If you care whether ChatGPT or Perplexity ever surfaces your work, this is one of the highest-value places for it to be mentioned. (Those systems already lean toward work that’s easy to find, which I wrote about in LLMs sharpen the Matthew effect.)

So why not just add the citation yourself?

Because Wikipedia treats that as a conflict of interest, and it does not go unnoticed. The relevant guideline, WP:SELFCITE, doesn’t flatly ban citing your own work, but it discourages it: keep it to a minimum, keep it in the third person, never give it undue weight, and when in doubt propose the edit on the talk page instead of making it yourself. Piling references to your own papers into an article is treated as a form of spam.

Two things make this a bad bet even when your work genuinely belongs there. First, every edit is public and permanently tied to your account in the page history, so “I added a citation to my own paper” stays on the record. Second, experienced editors patrol new citations and revert anything that reads as promotion. So the move that would do the most for your AI visibility is one you’re, by policy, not allowed to make directly. That is the whole problem. The rest of this post is how to get the result without breaking the rule.

What actually makes an editor reach for your work?

Editors cite sources that are easy to trust and easy to use. You make your work that source before anyone needs it:

• State the finding in one quotable sentence. An editor writing “studies have found X” needs a source that says X plainly. If your result is buried in hedged prose, it’s easier to cite the paper that said it clearly. (Same discipline as writing so AI search can read you.)
• Give it a stable, open link. A DOI that resolves, and a version anyone can actually read. Verifiability is the whole culture of the place; a paywalled PDF that a volunteer cannot open is a source they will skip.
• Earn secondary coverage. Wikipedia leans on secondary sources, a news write-up, a review article, a press piece that got picked up, more than on the primary paper by itself. Coverage is often what pulls a finding into an editor’s view in the first place, which is why a press release that actually lands does double duty here.

None of this is a trick. It’s the same thing that makes your work citable anywhere: a clear claim, a link that works, and someone other than you having noticed.

Can I touch a Wikipedia article in my own field at all?

Yes, and this is the version of it that works. Find an article in your area that’s thin or badly sourced, an unsupported claim, a “citation needed” tag, a statement propped up by a weak reference, and fix the sourcing with the best source available. Most of the time that source won’t be yours, and that’s fine. You’re improving the article, which is the point, and you’re learning where your field’s coverage is actually thin.

If your own work genuinely is the best source for a claim and nothing else fits, still don’t just insert it. Go to the article’s talk page, propose the specific edit, disclose that it’s your work, and let an uninvolved editor decide. That’s exactly what WP:SELFCITE tells you to do, and an edit that an independent editor accepts is one that sticks.

Where Loud Camel fits

This is slow, boundary-respecting work, and most of it’s knowing where your field’s conversation already points and where the sourcing is thin. That’s the kind of thing Loud Camel watches for you. It tracks where your work is already mentioned across the open web, flags where secondary coverage is missing (the gap that keeps you out of an editor’s reach), and drafts the outreach when there’s a real person worth contacting. It drafts; it never sends on its own. You stay the one who decides.

Common questions

What is the single best first step? Make one finding quotable and openly readable, then earn one piece of secondary coverage. That’s what moves your work into the pool editors actually draw from.

How long does this take to pay off? Longer than an edit you could make in a minute. That is the trade: the fast version gets reverted, the slow version stays.

The honest bottom line

The uncomfortable part is that the step that would help most here is one you’re not allowed to take yourself, so the work is indirect: be citable, be open, be covered, and let someone else make the edit. I could be wrong about how quickly the AI-visibility payoff arrives, since that part is still moving. The conflict-of-interest part is not an opinion, though. It’s policy, and the fastest way to get your citation deleted is to add it yourself.