The second most viewed page on Wikipedia in the past week was its entry on the blockbuster Dhurandhar 2. The excitement was the result of media coverage of a tussle between editors of the entry over whether to label the film “propaganda”. So intense was the “edit war” that Wikipedia’s founder, Jimmy Wales, was compelled to remind the editors in favour of describing the film as “propaganda” that the website is meant to uphold a “neutral point of view”. When challenged, Wales conceded to the disgruntled editors that they could note that “reliable sources” had described the film as propaganda. Thereupon, these editors began copiously citing reviews in Left-wing publications that described the film as “propaganda”. Consequently, the term “propaganda” now appears over 25 times in the entry on Dhurandhar 2. By contrast, the term appears three times on the page for Wolf Warrior 2 and only once for Top Gun: Maverick — films that were actively endorsed by governments in China and America.
The Dhurandhar 2 imbroglio underscores serious deficiencies in the Wikipedia enterprise. The first is blindness to self-interest. Wikipedia prides itself on being a “free encyclopedia”. What this really means is that its editors perform unpaid labour. Now, unless Wikipedia has miraculously solved the eternal question of how to make humans genuinely selfless, we must assume that editors who devote countless hours to writing, editing, and disputing entries on “contentious” topics have a strong interest in them — in other words, they are likely ideologues or partisans.
This worry is greatly exacerbated by Wikipedia’s willingness to let editors use aliases. This means we must believe that Wikipedia’s entries are produced by individuals who, unlike even saints, want no recognition for their charitable work. This is not the image that comes to mind, however, when one visits the clunky “Talk” section that underpins every entry on Wikipedia. Here, it becomes apparent that the decision-making process as to what is included in an entry can be rather arbitrary. The acronym-heavy discussions are dominated by “experienced” editors who can decide what the “consensus” view on a “contentious” topic is and thereby “close” discussions in favour of their preferred point of view.
This brings us to a third shortcoming. Wikipedia’s leaders are conscious of the threat of bias (as Wales’ intervention shows). The problem is that they try to combat bias through procedural rules. This is a mirage because rules have to be interpreted and applied. Thus, for example, editors who wanted to label Dhurandhar 2 as “propaganda” worked around the “neutral point of view” rule that Wales stressed by citing another rule related to “due weight”, which instructs editors to prioritise “reliable sources”. In other words, the way to control the narrative in a Wikipedia entry is to cite sources favourable to one’s interpretation as “reliable” (and to discredit as “unreliable” those that offer a contrary perspective), and to then “canvass” support among sympathetic editors, who then “protect” the entry from “disruptive editing” by challengers. The end result is not a marketplace of ideas but competition among cartels of cranks.
Wikipedia’s leaders have responded to this difficulty by providing a series of “dispute resolution” mechanisms crowned by the Arbitration Committee (ArbCom). But this approach is not designed to yield substantive remedies. The members of the ArbCom, who also use aliases, focus purely on matters of procedure and conduct. They can, for example, “indefinitely” ban editors who egregiously violate Wikipedia’s rules (though this does not prevent these individuals from resuming their old ways under a new alias, a technique known as “sock puppetry”). This leaves untouched the more subtle way of pushing a favoured narrative, which is to emphasise the “reliability” of one’s favoured sources and to demand that “due weight” be accorded to them. When such substantive judgment is exercised pseudonymously, it becomes difficult to detect whether an editor or indeed a clique is behaving maliciously. This is precisely why we expect courtrooms to operate publicly: It allows observers to evaluate biases and hold judges responsible for foul judgement.
This deficiency raises a broader concern. What happens on Wikipedia no longer stays there. Because Wikipedia contains free and structured data, it has a disproportionate impact on the training imparted to Artificial Intelligence (AI) models. This means that the biases of its editors are likely to be propagated widely. It is, therefore, insufficient to say that those who use Wikipedia do so at their own peril. Wikipedia defends itself against such criticism by pointing to the long-term advantages of “radical transparency”. There are indeed studies that show that entries on Wikipedia tend to become less partisan in the long run, as new editors enter the fray and challenge obvious biases in existing entries. But this does not really help when Wikipedia and AI are used — as they commonly are — to quickly come to grips with current events and controversies. In such cases, the reader does not obtain the “wisdom of the crowd” but the self-interested views of advocates who are better organised and resourced, with the “truth” lagging behind as editors squabble and manoeuvre on the Stone Age interface of the “Talk” page.
The criticism voiced here is not meant to imply that Wikipedia is a malign force. Its great value lies in the fact that it challenges credentialism and allows every individual to contribute to collective knowledge. Unfortunately, its rules have turned the great gift of freedom against itself. Rather than defeat gatekeeping, Wikipedia encourages a particularly invidious form of it.
The obvious remedy to this deficiency is to introduce a tincture of responsibility by requiring editors to verify and disclose their identity — on the grounds that those who wish to contribute to a collective enterprise that deeply affects public discourse ought to do so openly. This course of action is, admittedly, not costless. The prospect of public criticism may leave some individuals unwilling to contribute to Wikipedia. For instance, an individual living in West Bengal may not feel free to openly contribute to the entry on the Trinamool Congress. Even so, the burden being imposed on such an individual is hardly unreasonable. Those who greatly value their privacy or security are not being denied a chance to speak; they can continue to write pseudonymously on other forums such as Substack or X. Nor should we forget the other side of the ledger — there is a very real danger that permitting aliases in order to protect speech can end up endangering Wikipedia’s mission. It will, as recent events show, lead to Wikipedia being seen not as a repository of knowledge but of propaganda.
Rahul Sagar is Global Network associate professor at NYU Abu Dhabi. His recent books include The Progressive Maharaja and To Raise A Fallen People. The views expressed are personal
