WHEN KENNETH LAY, the disgraced Enron chief, died suddenly this month, Reuters conducted an important experiment: the news service carefully monitored the way the death was described on Wikipedia, the online encyclopaedia. Wikipedia articles can be written and edited by anyone, on every subject, at any time. Launched just five years ago, Wikipedia is free, vast and evolving at an astonishing pace, with some 4.6 million entries at the last count. For many people, Wikipedia has become the first port of call for information.Increasingly, it is the engine creating instant, modern knowledge.

Lay’s abrupt death, six weeks after his fraud conviction, demonstrated that online knowledge can evolve in mutant patterns. The news broke at 10am. Over the next half-hour, Lay’s entry on Wikipedia was edited five times. At 10.06 the death was “an apparent suicide”; two minutes later it was “heart attack or suicide”; at 10.11, the first moralist chimed in with a declaration that “the guilt of ruining so many lives finaly (sic) led him to suicide”; a few minutes later “a doctor”, pronounced that the “stress” of Lay’s trial had probably killed him. Thus, in the space of a few minutes, Wikipedia had produced four different causes of death: suicide, heart failure, nervous strain and the retribution of a just God.

For Wikipedia’s detractors, the incident was further proof of the dangers inherent in a DIY encyclopaedia. Wiki-wiki is Hawaiian for “quick”, but critics point out that Wikipedia is so fast that it also tends to be sloppy, partial and inaccurate, sometimes wildly so. About 13,000 people actively contribute to Wikipedia with articles and edits: some of these are experts; but some, inevitably, are nutters.

Wikipedia’s defenders, however, argue that the death of Lay shows the resilience of the system. True, there were initial inaccuracies, but these were edited out over time. The system worked precisely as intended.

The incident was only the latest skirmish in a growing war between the new Wikipedia and the older forms of reference. This is more than a conflict between new technology and a more academic approach, between paper information and digital information. At stake is the nature of knowledge itself, and truth.

The internet had evolved a new form of information, a shallow, broad, fast, patchy and extremely useful reservoir that should be absorbed with caution and used only for specific purposes. Wikpedia has the same relationship with an encyclopaedia that yesterday’s news reporting has with tomorrow’s history book. Wikipedia is a first draft. It is not truth. But so long as it is understood and used in that way, it may prove to be one of the most spectacular inventions of the 21st century.

The online encyclopaedia was a simple, brilliant idea, the latest flowering of the Enlightenment ideal of the collective pursuit of truth. By pooling our collective knowledge, gradually weeding out the mistakes and the myths, we would arrive at a “repository of knowledge to rival the ancient library of Alexandria”, a fantastic, free experiment in intellectual democracy.

It has not quite worked out that way. Mistakes and libels crept in, through accident or malice. Entries were vandalised, others tweaked by the subjects themselves to paint a more flattering picture. In the entry on McDonald Corporation, a link to a book attacking the fast-food chain mysteriously disappeared, replaced by an anodyne volume on the company’s history.

Any trawl of Wikipedia reveals that this is not a democracy, but a self-selecting oligarchy. The people who have the time and inclination to write and edit Wikipedia tend to be younger, and interested in pop culture or technology. The space devoted to the glamour model Jordan’s breast implants is as long as the entire entry for the Yi language, spoken by 6.6 million Chinese.

There is also a danger of what some critics call “online collectivism”, or “digital Maoism”. Just because a majority of people happen to believe something does not authenticate it. History is littered with unpleasant examples of moments when the collective voice, ignorant or misled, has drowned out dissent. Wikipedia gropes towards a consensus, but that is very different from truth.

All these criticisms are valid, but they miss two crucial points. Wikipedia is not an authoritative research tool, but a work in progress that will never, by definition, be finished. It makes no claim to omniscience; indeed, it cautions strenuously against treating any entry as valid. It should be used as a rough guide to knowledge, as a starting point for intellectual exploration.

Purists who fear that a cluttered glut of half-baked information is being foisted on to impressionable minds give too little credit to the online public. Since Wikipedia users know they are also the creators, scepticism is built into the process. Most people approach Wikipedia with a firm understanding of its limitations; any students who cut and paste Wikipedia errors into their work do not deserve to be in higher education.

The second point is that Wikipedia is here to stay, gradually improving, and growing more influential by the minute. If there is a subject that you know and care about, then it is becoming an intellectual duty to ensure that the entry on Wikipedia is as accurate as possible.

Many years ago, I wrote a book about a Victorian crook called Adam Worth, a subject so obscure that no one had ever written a book about him before, or since. When I found an entry for Worth on Wikipedia, I was at first astonished, then flattered to find the book cited in the references, and then slightly infuriated: whoever had written the entry had plainly read my book and summarised it, but added several small but irritating errors.

At first, I ignored the mistakes. This was only Wikipedia, after all. But the landscape of knowledge has changed since then, and I have joined the club. Wikipedia should always be taken with a pinch of salt. But the more we contribute and revise, the less salt we will need. We are all Wikipedists now.

Times Recommends

• It’s the beginning of the affair
• Afghanistan: still a price worth paying
• Heroes of the downtrodden