How Wikipedia actually works

Most people use Wikipedia every day and have no idea what happens behind the articles. They assume it's either a free-for-all where anyone writes whatever they want, or a tightly controlled operation run by experts. Neither is right. The reality is weirder and more interesting than either caricature.

Wikipedia is a bureaucracy. A functional, volunteer-run, somewhat anarchic bureaucracy with its own courts, its own police force, and its own legislature. It has survived for 25 years not because of some magical faith in human nature, but because of a set of well-designed (if sometimes maddening) governance structures.

Let me walk you through how the machine actually runs.

Anyone can edit (but here's what actually happens)

Yes, anyone can click "Edit" on almost any Wikipedia article and change the text. This is still true in 2026. You do not need an account. You do not need permission. You just click and type.

What happens next depends on the article.

For low-traffic articles with no special protections, your edit goes live immediately. Nobody reviews it first. If you add a sentence to the article about a small town in Nebraska, it's visible to the world within seconds. This is how most Wikipedia edits work, and it's the part that makes people nervous.

But for articles about controversial topics, living people, or subjects prone to vandalism, the process is different. These articles have various levels of "protection" that restrict who can edit them. The article about the current U.S. president, for example, can typically only be edited by accounts that are at least 30 days old with at least 500 edits. Some articles are fully locked and can only be edited by administrators.

Even for unprotected articles, every edit is watched. Wikipedia maintains a "Recent Changes" feed that scrolls constantly, showing every edit made across the entire encyclopedia in real time. Volunteer patrollers monitor this feed. Automated bots monitor it too. If you add something obviously wrong or malicious, it's usually reverted within minutes. Often within seconds.

The hierarchy nobody talks about

Wikipedia's egalitarian reputation is misleading. There is a clear power structure, and it matters.

Regular editors can edit unprotected articles and participate in discussions. That's about it.

Autoconfirmed users have accounts at least 4 days old with at least 10 edits. They can edit semi-protected articles and create new articles. This is the first real threshold, and it exists to stop drive-by vandals.

Extended confirmed users have accounts at least 30 days old with 500+ edits. They can edit articles under "extended confirmed protection," which covers many controversial topics.

Rollbackers have the ability to instantly undo the last set of edits by a single user with one click. This is a tool for fighting vandalism, and it's granted by administrators to trusted editors who request it.

Administrators (also called "admins" or "sysops") are the backbone of Wikipedia's governance. There are roughly 1,100 active administrators on English Wikipedia. They can delete pages, block users, protect articles, and access deleted content. Becoming an admin requires surviving a community vote called RfA (Requests for Adminship) where your entire editing history is scrutinized. Passing rates have dropped over the years, and most candidates now have thousands of edits and years of experience before they even attempt it. The last time I checked, fewer than 50% of candidates succeed.

Bureaucrats can grant and revoke admin status. There are only about 12 active bureaucrats on English Wikipedia.

Stewards operate across all Wikimedia projects (not just English Wikipedia). They can grant any permission on any wiki. There are around 30 stewards globally, elected annually by the entire Wikimedia community.

CheckUsers can view the IP addresses behind accounts. This power is tightly controlled because it has privacy implications. There are fewer than 40 CheckUsers on English Wikipedia.

The Arbitration Committee (ArbCom) is Wikipedia's supreme court. It handles disputes that can't be resolved through normal discussion. ArbCom members are elected by the editing community, serve terms of varying lengths, and can issue binding decisions including topic bans and site bans. They handle maybe 10-20 cases per year, but those cases tend to involve the most contentious areas of the encyclopedia.

Role	Requirements	Key powers	Approximate count
Regular editor	None (no account needed)	Edit unprotected articles	Millions of IPs
Autoconfirmed	4 days, 10 edits	Edit semi-protected, create articles	~300,000+
Extended confirmed	30 days, 500 edits	Edit extended-protected articles	~30,000+
Rollbacker	Granted by admin	One-click mass revert	~3,500
Administrator	Community vote (RfA)	Delete, block, protect	~1,100 active
Bureaucrat	Community vote	Grant/revoke admin	~12 active
CheckUser	ArbCom appointment	View user IP addresses	<40
Steward	Global community vote	Any permission, any wiki	~30 globally
ArbCom member	Community election	Binding dispute resolution	15 per term

The five pillars (and the three policies that actually matter)

Wikipedia has five foundational principles called "The Five Pillars." In practice, three specific policies do most of the heavy lifting in day-to-day editing disputes.

Neutral Point of View (NPOV). Articles must represent all significant viewpoints fairly and without editorial bias. This doesn't mean giving equal weight to all views. The flat earth theory doesn't get equal space in the article about Earth. "Neutral" means proportional to the reliable sources available.

NPOV disputes are the most common content arguments on Wikipedia. Editors fight endlessly about whether a particular phrasing gives too much or too little weight to a perspective. These fights are especially intense on political, religious, and scientific topics where different communities have different views about what "neutral" means.

Verifiability (WP:V). Everything in Wikipedia must be attributable to a reliable, published source. It doesn't matter if something is true. If you can't cite a source for it, it doesn't belong in Wikipedia.

This policy is more radical than it sounds. It means Wikipedia isn't trying to describe reality. It's trying to describe what reliable sources say about reality. The distinction matters. If a fact is true but no reliable source has published it, Wikipedia won't include it. If two reliable sources disagree, Wikipedia presents both views. The encyclopedia is a mirror of published knowledge, not of truth itself.

No original research (WP:NOR). You can't use Wikipedia to publish your own theories, analyses, or conclusions. Even if you're the world's leading expert on a topic, you can't add unpublished findings to the article. You need to publish them somewhere else first, and then someone (preferably not you) can cite that publication in Wikipedia.

This policy drives some academics absolutely crazy. They see errors in their field's Wikipedia articles, they know the correct information, and they're told they can't add it because they haven't published it in a citable source yet. The policy exists because Wikipedia can't verify expertise. Anyone claiming to be an expert might be. Or might not be. Requiring published sources is the imperfect solution to an impossible problem.

How articles get deleted

Creating a Wikipedia article is easy. Keeping it alive is harder. Wikipedia has a formal deletion process called "Articles for Deletion" (AfD) where editors debate whether an article should exist.

The most common reason for deletion is "notability." Wikipedia requires that article subjects have received significant coverage in reliable sources independent of the subject itself. Your local pizza shop, no matter how good it is, probably doesn't qualify. Neither does your band, your startup, or your self-published novel, unless external sources have written substantially about them.

AfD discussions run for seven days. Any editor can participate, and a closing administrator weighs the arguments (not just the votes) to determine consensus. "Delete" and "Keep" are the main outcomes, but there's also "Merge" (combine with another article), "Redirect" (turn into a redirect to a related article), and "No consensus" (article stays, but another AfD can be filed later).

About 50-60% of articles that reach AfD end up deleted. But many more articles are deleted through a faster process called "speedy deletion" for obvious cases: copyright violations, nonsense, blatant advertising, hoaxes, and articles about subjects with no conceivable claim to notability.

Featured articles: the gold standard

Not all Wikipedia articles are created equal. At the top of the quality hierarchy sit "Featured Articles" (FAs), marked with a small bronze star in the corner. These represent Wikipedia at its absolute best.

To earn Featured Article status, an article must pass a rigorous peer review process. The criteria are demanding: comprehensive coverage of the topic, impeccable sourcing with inline citations for all significant claims, neutral and well-written prose, a stable version with no ongoing edit wars, appropriate images with proper licensing, and a proper lead section that summarizes the entire article.

The average Featured Article takes over two years to develop from initial creation to FA status. Some take much longer. The article on the city of Melbourne went through 14 major revisions over three years before passing its FA review.

As of early 2026, English Wikipedia has about 6,400 Featured Articles out of roughly 6.8 million total articles. That's less than 0.1%. The bar is high, and the community of FA reviewers is small and meticulous.

Below Featured Articles are "Good Articles" (about 40,000 of them), which meet a slightly less demanding set of criteria. Together, FAs and GAs represent the best-sourced, best-written content on Wikipedia, and they're still a tiny fraction of the total.

How edit wars get resolved

When two or more editors disagree about the content of an article and keep reverting each other's changes, that's an edit war. Wikipedia has a specific policy against it: the "three-revert rule" (3RR) says that reverting another editor's changes more than three times in 24 hours can result in a block.

But the three-revert rule is a blunt instrument. Real content disputes are resolved through talk page discussion. Every Wikipedia article has an associated talk page where editors can argue about content, propose changes, and try to reach consensus. For contentious topics, these talk pages can be enormous. The talk page archives for the article on the Arab-Israeli conflict contain millions of words of debate spanning two decades.

If talk page discussion fails, editors can escalate to various dispute resolution processes. "Third opinion" (3O) brings in an uninvolved editor to give their view. "Requests for Comment" (RfC) opens the question to the broader community. Mediation brings in a trained mediator. And for the most intractable disputes, the Arbitration Committee can impose binding solutions, including banning specific editors from specific topic areas.

The process is slow. It's often frustrating. Experienced Wikipedians sometimes joke that Wikipedia's dispute resolution is designed to exhaust everyone until they stop caring. But it mostly works. Eventually, most articles reach a stable state that, if not everyone's first choice, is at least acceptable to the active editors.

The editing ecosystem

Wikipedia doesn't exist in isolation. Around the core editing community, there's an ecosystem of tools, projects, and initiatives.

WikiProjects are groups of editors focused on specific topic areas (WikiProject Medicine, WikiProject Military History, WikiProject Chemistry, etc.). They coordinate efforts, maintain quality standards for their area, and help recruit new editors. WikiProject Medicine, for example, has been particularly successful at improving the quality of health-related articles, working with medical professionals to ensure accuracy.

Bots do an enormous amount of work on Wikipedia. They fix formatting errors, update statistics, patrol for vandalism, tag articles for issues, and perform thousands of other maintenance tasks. Some bots make millions of edits per year. The bot approval process is rigorous because a poorly designed bot could damage thousands of articles before anyone noticed.

Wikimedia Foundation is the nonprofit that operates the servers and employs about 700 staff. The Foundation does not control Wikipedia's content. That distinction is important and frequently misunderstood. The Foundation handles technology, fundraising, legal issues, and partnerships. Editorial decisions are made by the volunteer community.

Why this matters for Bluffpedia

If you understand how Wikipedia works, you understand why certain patterns appear in real articles. The slightly uneven prose, the careful hedging around controversial claims, the specific citation patterns, the way articles about living people are more carefully worded than articles about historical events. These are all artifacts of Wikipedia's governance structure and policies.

When you play Bluffpedia, you're implicitly testing your understanding of these patterns. AI-generated summaries mimic Wikipedia's surface style but miss the deeper structural features that emerge from the collaborative editing process. Real articles have the fingerprints of their creation process embedded in them, if you know what to look for.

The volunteer problem

I want to end with something that doesn't get enough attention. Wikipedia is maintained by a shrinking pool of volunteers. Active editor counts on English Wikipedia peaked around 2007 at over 51,000 and have declined to roughly 30,000-35,000 (though the 120,000 figure for all Wikimedia projects includes editors across all languages).

This matters because the encyclopedia keeps growing while the workforce maintaining it shrinks. Articles that were well-maintained in 2010 may not be as carefully watched in 2026. Vandalism that would have been caught in seconds might linger for hours on less-trafficked articles.

The community has been discussing this problem for years without finding a solution. Editing Wikipedia requires a significant time investment, the interface has a steep learning curve, and the social dynamics of the editing community can be unwelcoming to newcomers. Multiple initiatives to recruit new editors from underrepresented groups have had mixed results.

Wikipedia is one of the most important information resources humanity has ever built. It's maintained by a group of volunteers roughly the size of a mid-sized town. That's both inspiring and a little terrifying. The next time you read a Wikipedia article and find it well-written, accurate, and properly sourced, take a second to appreciate the hundreds of hours of human labor that went into making it that way.