Wikipedia, one of the world's 15 most popular websites, has more than 10 times the number of entries in Encyclopaedia Britannica

Related articles

• Stephen Hawking endorses colonisation of the Moon
• Colossal squid dwarfs giants
• U.S. military deems homosexuality a 'mental disorder'
• Paralysed rats cured; hope seen for humans
• Breath test for cancer

A car that runs on water, a new form of energy derived from 'hydrinos', a 'cognitive-theoretic' model of the universe. They sound like fantastical concepts you might find in the pages of the latest Harry Potter book, or Alastair Reynolds novel.

In fact, they're all entries in the wildly popular collaborative online encyclopaedia, Wikipedia. At first glance, the site's main entry for 'hydrino theory' looks objective and reliable; but link to the entry's discussion section – where authors and editors can debate content – and you'll find another side to the story. The discussion reveals the colourful history and heated debate behind the entry's evolution. So how many other Wikipedia entries are contested?

The avowedly populist online encyclopaedia, created by volunteers from across the world, is among the world's 15 most popular web sites and has 40 million readers in the U.S. alone. And despite its collaborative nature, many academics are happy for their students to use the resource.

"I'm a big fan of Wikipedia," says Sean Carroll, a physicist with the California Institute of Technology in Pasadena. "It evolves rapidly in response to current events, can grow to arbitrary size, and contains links to interesting outside resources – what more could you want?."

A recent in-depth piece in the New Yorker credited Wikipedia for its 'free-information-for-all' ethos, the all-inclusiveness of its varied articles and its vast number of entries – currently more than 10 times the number of entries in Encyclopaedia Britannica.

However, as the populist encyclopaedia has grown, so has the question of its reliability. An increasing number of academics - many of whom have helped edit the resource to maintain informal quality control - are concerned that Wikipedia is becoming a stronghold for cranks: people who anonymously submit and edit entries on pet subjects to bolster the credibility of highly questionable theories.

The hydrino theory listed above is just one example of crank science trying to gain credibility through the online encyclopaedia. But there are hundreds of others, with legions of supporters willing to donate their time and - often anonymously - promote questionable science.

A 'Cognitive Theoretic Model of the Universe', which claims to be a metaphysical theory of the relationship between mind and reality, no longer has its own entry on Wikipedia, but the subject has generated a great deal of heat on the discussion page for its creator, one Christopher Michael Langan.

Two supporters of Langan, with the online pseudonyms Asmodeus and DrL, have been permanently banned from editing the article about Langan because of their overtly one-sided support of his theory. One Wikipedia editor, who asked not to be identified, said there was reason to believe that Asmodeus was Langan himself, and DrL, his wife.

Controversy such as this has fuelled concern over the accuracy of Wikipedia entries in general. In April 2007 the history department of Middlebury College in Vermont, USA, banned its students from referencing any material on Wikipedia. And this might only be the tip of the iceberg.

Because Wikipedia allows anonymous editing as a matter of principle, any promoter of a crackpot idea can potentially sign-in and add comments and links supporting baseless theories. This is called 'wikishilling' in Wikipedia lingo.

One way self-promoters operate is by creating 'sock puppets': editing their pet pages anonymously from internet addresses different to the ones they normally use at work or home. This makes their edits and entries appear to be the work of other contributors.

More narcissistic cranks use Wikipedia's perceived reliability to promote themselves or their personal cause. They typically create a user account under a pseudonym, then author an entry about themselves. They – or their friends and associates - can then create supplementary articles on Wikipedia to support the original entry.

For ideas like hydrino theory, which have the potential to generate serious financial investment, Wikipedia can be a way of conferring legitimacy on something that wouldn't stand up to scientific peer review. Proposed by medical student Randell Mills, the hydrino theory suggests that large amounts of energy could be tapped when hydrogen atoms make a transition from their ground state of energy to a hitherto unknown state, supposedly below this level.

According to Robert Park, a physicist at the University of Maryland, USA, Mills was even able to persuade NASA to take his idea seriously. The U.S. space agency was convinced enough to purchase some cold-fusion-cell-like devices and test Mills's prediction, with "inconclusive" results.

Park keeps an eye on the hydrino theory from his blog. "I have mentioned Mills and hydrinos several times," he says. "Each year … I put out my 'March Madness' issue, which began with the 'discovery' of cold fusion by Pons and Fleischmann on March 23, 1989."

He is not surprised that Mills and others with questionable theories have set up camp in Wikipedia. "This, of course, is what everyone predicted would happen to Wikipedia. Too bad. Science owes its success and credibility to openness. Wikipedia, it was hoped, would simply take openness to a new level. This is a level too far. Maybe we can work through this trend," he says.

So what can the army of editors maintaining Wikipedia do to improve its credibility? The arbitration committee set up to provide a review of ongoing controversies is starting to formally address the issue (see science apologists and pseudoscience, for example).

However, a major stumbling block to catching cranks is the fact that the tools editors must use to uncover wikishilling or violations of the rules are also a threat to privacy. They rely on a level of snooping - tracing IP addresses, for example, to the location where users log in - that many Wikepedians find ethically questionable. And there's no consensus among Wikipedians that stricter policing and oversight are appropriate, given the populist goals of the online encyclopaedia.

Wikipedia now seems caught between a rock and a hard place. If it stays true to the populist and libertarian ideals of its co- founder Jimmy Wales, it may eventually sink under the weight of unreliability. But if it demands more accountability from contributors, it will lose the anti-elitist ideals that inspired its creation.

Readers' comments