IBM Employees Caught Editing Wikipedia

An anonymous reader writes "Corporate employees editing Wikipedia articles about themselves or their employers sometimes commit major violations of Wikipedia's "bright line" against paid editing, devised by Jimbo Wales himself, to prevent 'COI' editing. (Consider the recent flap over the firm Wiki-PR's activities, for example.) Yet the Wikipediocracy website, run by critics of Wikipedia management, has just published an article about IBM employees editing Wikipedia articles. Not only is such editing apparently commonplace, it's being badly done as well. And most bizarrely, one of the IBM employees is a Wikipedia administrator, who is married to another Wikipedia administrator. She works on the Watson project, which uses online databases to build its AI system....including the full text of Wikipedia." Reading about edit wars is also far more informative (if less entertaining) than reading the edit wars themselves.

Comments

CN

[OptimalCynic]

"(if less entertaining)" [citation needed]

TheAgriculture Ministry is not in charge of Gundam

[cheesybagel]

A lot of people browse the web while they are working. I can understand that there could be a conflict of interest when people are editing topics which are biographies of themselves or their own employers. But as long as they are not deleting facts they don't like and are adding actual information that other people may not have easy access to I do not see how that is a problem.

Re:TheAgriculture Ministry is not in charge of Gun

[Anonymous Coward]

But as long as they are not deleting facts they don't like...

That's exactly what they're doing.

Wikipedia is utterly broken anyway.

[Anonymous Coward]

What matters is transparency. You can't prohibit people with bias from editing the truth in a "truth by democracy" project - you can only hint strongly to them that they need to better hide their true identity. But you could, if you wanted to put an ounce of scholarly rigour into Wikipedia, make it so that people reveal their biases. There is nothing wrong with IBM employees contributing toward an article on IBM, as long as everyone knows that the perspective is that of an IBM employee - similarly, there's nothing wrong with someone who has invested time and emotion into some political view or war or comic book, as long as they are clear on their opinions.

The first worst thing about Wikipedia is that editing it is about a tenth as productive as editing just about any other online resource, because you have to continually fight to maintain high standards. The second worst thing is that it tries to pretend that you can eliminate biased people, rather than acknowledge that bias exists and tackle how to be open about it.

Re:Wikipedia is utterly broken anyway.

[AthanasiusKircher]

So long as what people are adding is verifiable, who cares who they are?

Well, in my opinion, the only way to really fix Wikipedia would be to allow expert help -- and in many cases, that may actually benefit from having someone with intimate knowledge of something.

Unfortunately, Wikipedia policies discourage experts from contributing (sometimes by official policy, sometimes just by the attitude of frequent editors). It's really more like the Wild West... with some bizarre lawyer-like class running everything.

The only way to bias a Wikipedia article while remaining verifiable is to delete stuff, and that's where the wars tend to happen.

That's not true at all. The "verifiability" requirement is probably one of the most broken elements of Wikipedia, next to the "notability" criterion. Not that facts shouldn't be verifiable: of course they should be, and of course reputable sources should be used. However,

Verifiable != True

Nor should we think that verifiability is some sort of useful proxy for truth. There an important reason why people tend to swear in legal proceedings to "tell the truth, the whole truth, and nothing but the truth."

Stating "facts" without context is useless. I can tell you that the "deadly chemical" X has been found in a number of snack foods, for example. But what does that mean? Normally, people don't actually mention things unless they are notable, so the fact that I mention this seems to mean that there must be a greater significance. It seems to imply that someone put that chemical in snack foods, or that the manufacturer used contaminated ingredients, or that the manufacturer didn't do adequate screening and quality testing, or something else bad.

But what if I now add the facts: "Chemical X occurs naturally in soil, groundwater, and most food items at a rate of 10 ppb. In the snack foods in question, chemical X had a concentration of 0.02 ppb."

Now, all of those implications seem way off-base, no? Not only did we make an incorrect assumption from a true statement, but we actually assumed the opposite of what is true: apparently, whatever the manufacturer is doing, they are actually producing a safer-than-average product. But the concentration is still non-zero, as it is in almost all food products.

You see this sort of thing on Wikipedia all the time from people with agendas, and from people who are just ignorant of the larger context.

Particularly in low-profile articles on obscure topics (like the humanities), you'll often see citations and quotations from scholarship that is 50 years old and from a book that isn't even on the topic of the article. Yes, it was "published" in a "reliable source" perhaps even in a book by a major university press, so it meets quite high standards of "verifiability," but it's not particularly representative of scholarship or what most people know to be true.

Someone with an agenda on a topic can really skew things this way. An article has citations to 5 scientific articles published in credible journals claiming X -- well, it looks like X is true. But almost every area of knowledge has some disagreement. What if there are actually 100 articles on the same topic that claim not-X, but they just don't happen to be in Wikipedia? The only person who becomes aware of this is the rare Wikipedia editor who does a search of the scholarly literature in some obscure field. Otherwise, articles can exist for years claiming things that are obviously not true. (In fact, given the propensity for scholars to exaggerate claims and implications in their own research, it's quite easy to even find 5 articles that don't even have data to prove X, but nevertheless assert X to be true in their discussion sections... while there might be 25 articles actually on topic out there which refute X.)

So yeah, there are loads of ways to skew an article by providing "verifiable facts" from "v

Funny idea

[jones_supa]

I just got this crazy idea. You know those videos in YouTube where Hitler gets worked up about something and there's various fake subtitles people have crafted over that clip. Make one where Hitler discovers Slashdot Beta.

Re:Funny idea

[Anonymous Coward]

Those aren't fake, and Hitler is really pissed about Beta.

You can't prevent all bias.

[areusche]

There is only so much bias you can remove from certain articles. Especially when this is a massive crowd sourced project. In all honesty, I trust the large amount of people who frequent these articles to help keep astroturfing to a minimum. You'd think by now, people would be good at critically looking at a piece of information and being able to identify the bias of the writer.

On a completely random side note, the value wikipedia provides for FREE is immense. Trust me, it is one of the best resources we have and the citations can help me find other details and continue reading if I so wish. If I was forced to go back in time to the 1500s and could only take one thing, I'd take this http://www.good.is/posts/wikip...

Re:You can't prevent all bias.

[wonkey_monkey]

If I was forced to go back in time to the 1500s and could only take one thing, I'd take...

...a time machine.

Experts in subject

[schneidafunk]

I get how there's going to be bias, but I think it is OK for someone to write an article about themselves to start it. Or in a company's case, to correct factual errors or give a history of the company.

Wikipedia needs MORE paid editors

[hydrofix]

Just like with the Linux kernel, it's a high time the Wikipedia community gave up the futile resistance to paid editing. It's already happening, and denying it is only embarrassing with "revelations" like this IBM case. What goes to the whole Wiki-PR debacle, turns out all the company was doing was correcting errors, libel and defamation that anonymous Wikipedia editors hiding behind pseudonyms and IP addresses have been adding to Wikipedia.

As it stands, Wikipedia is essentially an anarchy where anyone can publish all sorts of lies and propaganda, and companies like Wiki-PR are needed so that those, who are damaged by misinformation that anonymous Wikipedia editors publish, can hire neutral editors to fight the anonymous hoaxers. Wikipedia's own volunteer community has been since long overwhelmed by the sheer amounts of vandalism and biased information added every minute, and only the most obvious cases of misinformation and fraud are ever caught. But instead of celebrating the work that Wiki-PR was doing for the people and companies who have fallen victim to the terror of Wikipedia misinformation, the company behind Wikipedia instead chose to demonized Wiki-PR to media and threatened to sue them.

What's really worrying, is that Google gives Wikipedia a "boost" in its search rankings. So for example, any hoaxer can easily use Wikipedia to publish misinformation about people, products and companies that they don't like. Then anyone searching Google for the name of the person, product or company are immediately served the Wikipedia page on the subject. This page is often full of misinformation and propaganda, while those concerned (like the employees of the said company or the person being defamed himself) are forbidden from correcting the article. Previously, Wikipedia admins were satisfied with just banning those fighting the misinformation under the "conflict of interest" doctrine. But now, the company behind Wikipedia has demonstrated that they are ready to sue you if you want to correct the lies that are being distributed through their platform.

March of the 50-Cent Army

[Guppy]

So how about propaganda-style editing backed by PR operations with sovereign backing? For instance, articles involving China, where the 50-cent army runs rampant over the more obscure topics (in contrast, popular and well-known topics are usually well-defended, so only subtle alterations tend to get through).

It's not just Wikipedia -- they're likely present on any western media forums considered high-traffic or influential in the realms of policy (for instance, The Economist's comment sections), where they crap up threads and start flamewars to disrupt topics critical of the PRC. It's hard to distinguish them from posters which may merely be jingoistic bozos, but their abundance and stubborn persistence is unusual, compared with topics about any other nation.

Edit, but disclose

[140Mandak262Jamuna]

Often they are in the best position and knowledge to contribute. Only thing I would ask for is disclosure so that others can watch and correct manipulation of pov.

Personally, I would not like to lose my freedom of expression to express my views on, say, finite element analysis or mesh generation just because I work for a company making commercial products in that area. But if I ever edit the wiki article of my employer, I would make damn sure everyone knows my background so that my biases conscious or unconscious are corrected.

Re:Edit, but disclose

[qbzzt]

This is pretty much IBM's policy. We're allowed to post about IBM, provided what we post is:

1. Non confidential (I can't tell you about the time machine we're building in the basement in Austin)
2. Not damaging (I am not allowed to be defamatory, for example - although I will say all of HP's employees who are black with yellow polka dots are habitual liars)
3. Clearly stated to be the opinion of a specific IBM employee, rather than the IBM corporation

It would be silly not to let IBM employees, for example, post about mainframes. A lot of the world's mainframe expertise is at IBM.

Disclaimer: I am an IBM employee, but these are my own opinions. They do not reflect the opinions of the IBM corporation in any way, shape, or form. Considering that IBM is not a natural person, I'm not sure if it even has opinions, beyond "money coming in good, money going out bad".

Re:Edit, but disclose

[Sarten-X]

I have long used the analogy that Wikipedia is like asking experts in a particular field. For the most part, you'll get accurate information representing the current state of the art. Sometimes, you'll stumble across an expert with bad data, or a conflict of interest, or another factor leading to him giving you incorrect information. You, as the reader, should expect this and seek verification of any questionable claims, but for obtaining a general understanding of an area of study, Wikipedia is good enough.

With that analogy in mind, why shouldn't you directly ask someone about their life and their work? Of course it will be biased, but it's more obvious than the random biases you'll encounter when asking others.

Re:Edit, but disclose

[Aighearach]

Unfortunately, most experts have jobs. Worse, their jobs are often in the same field they're an expert on!

Anything an expert might say is biased original research.

Wikipedia editing is only for out-of-work journalists, school children, and neckbeards. Nobody else can be trusted.

Re:Edit, but disclose

[Trepidity]

That's also more or less the actual Wikipedia policy, despite what the summary implies. There is no prohibition on people who are paid money also editing Wikipedia. In fact Wikipedia actively encourages it in some cases, such as trying to recruit more museum and library staff to contribute to Wikipedia.

What is generally prohibited is: 1) taking money to write promotional articles or "clean" articles on behalf of their subject, like Wiki-PR was doing; or 2) making edits on an article where you have a clear COI, and failing to disclose your relationship to the subject of the article. In practice even #2 only really matters if they're biased edits that would cause someone to raise an eyebrow.

Re:Edit, but disclose

[Pharmboy]

COI isn't against policy, it is simply discouraged. I've participated in a number of policy discussions (as an admin) and tried to initiate a number of policy initiatives on the subject matter, but there is no consensus. Disclosure is a good idea, but in no way, shape or form is it required by policy. Knowingly adding bad material or inaccurate material (regardless of COI) is still prohibited.

provenance

[markhahn]

let people perform whatever edits they want, but track the provenance of *everything*. let readers select some function of provenance as a rendering option, with the default being provenance of a pretty high standard of quality and non-conflicted-interest. letting people attach endorsements or upvotes is a pretty valuable kind of metadata anyway.

the problem seems to be the very idea that wikipedia should present a single, canonical version. absolutes are only found in faith, not the real world...

Re:TheAgriculture Ministry is not in charge of Gun

[Ghostworks]

Except that's not what the article accuses them of. The article mainly accuses them of editing badly.

For those who didn't RTFA, here's the high points:
* IBM was huge in computing, so why is it so poorly represented (in terms of article count, total kB of text, and editing quality) on Wikipedia, the self-appointed online repository of all human knowledge?
* people at IBM seem to be editing IBM-related articles, but not in any kind of organized way. (The article actually chastises them for FAILING to have any kind of organized method.) Mostly it's people editing articles about themselves or things that they have worked on.
* The person who worked on the Watson project is and admin on Wikkipedia, married to another editor and edited Wikipedia articles while on the job for IBM. (Almost as if she were passionate about it or something... and working on a project where her computer barfed up nonsense when it parsed a really poorly written article....)
* the three shadiest things that they mention are 1) a guy who created an article about an IBM award/title he won; 2) an editing fight about the relevance of a book that linked IBM to the third Reich (which went through the usual Wikipedia channels and ended up in favor of keeping the article); and 3) The guy who started BASH.org (and who happens to be at IBM) arguing that the page was relevant and should be kept (again, usual Wikipedia channels, this time not in BASH.org's favor)

So basically what we have here are the notions that:
* even relatively obscure people probably shouldn't edit articles about themselves to avoid bias (which strikes me as silly for biasing things hard in the other direction)
* that IBM needs to tackle Wikipedia in an organized way to make up for the lack of interest by anyone outside the industry in preserving this huge chuck of history...
* unless it stays away altogether, because they already have a huge company history page on their website.
* and that IBM-ers should not touch the articles that they are most likely to have specific knowledge on...
* ...or for that matter any article, no matter what they happen to find odd if they found it while at work.

like most fights on and about Wikipedia, this is a tempest in a teapot by people who do a poor job articulating whether the collaborative encyclopedia of all human knowledge is actually suppose to be any of those things and why.