User:Doc glasgow/The BLP problem

From Wikipedia, the free encyclopedia
Jump to: navigation, search

Statistics in the essay are significantly out of date.

To solve or mitigate a problem, you must first define it. This page is a workshop with the intention of trying to answer: "What are the BLP problems?" Suggestions on the talk page are very welcome.

Am I talking bollocks? Am I missing stuff?

(Suggested changes on talk page please, but do feel free to copyedit)


Ways in which a biographical article can be harmful to the subject[edit]

By harmful I mean either legally or ethically: that which may cause unjustifiable and avoidable harm or distress to the subject.

1. Un-reverted nonsense[edit]

(Including bad-faith personal attacks, and patently obvious untruths).

Vandalism is probably the aspect most easily understood by Wikipedians, it includes articles which have been blanked, filled with nonsense, or include obvious abuse. Although this is perhaps the category on which most article patrolers focus, it is usually the least harmful to the subject. Contrary to popular opinion, abusive commentary ("tom smith is an asshole") or patent untruth that would deceive none ("Britney comes from mars" "Prince Philip is really a woman") are highly unlikely to be legally actionable and, although it may embarrass or annoy the subject, actually reflect worse on Wikipedia than they do on the target. Verdict – whilst the crime-prevention element of vandal slayers are obsessed with it, it is in fact mostly harmless.

2. False allegations[edit]

Here we are talking about the untrue, but credible, allegation inserted into an article. Some of these may be intended as vandalism, others may have more malicious intentions. Whatever the case, these may cause the Foundation legal concern (it is not the point of this examination to consider the Foundation's immunity). More importantly, they can be distressing to the subject, or indeed patently damaging to the reputation, career, or commercial interests of the individual. Whilst false allegations and rumours are the stuff of the internet, the fact that Wikipedia is a self-described "encyclopedia" which protests a commitment to factual accuracy, may lend credibility to untruths. Mirrors and google may perpetuate the lies even if it is removed from Wikipedia. Verdict – highly harmful.

3. Biased editing[edit]

Here we are talking about an article or section which may be factual at any given point, but presents a false or misleading overall picture. This is the hatchet job. Hatchet jobs are not always self-evident, since the tone that describes each item can appear neutral and well-referenced. However, undue weight may be given to minor negative facts, and/or positive counterbalance omitted entirely. "Nelson Mandela is a convicted South African terrorist, who spent many years in the criminal justice system, and, for many years, advocated violence." It may also include articles that extensively quote critics and criticism.

Although Wikipedia has a strong commitment to NPOV, certain aspects of Wikipedia can work in favour of the biased editor. Wikipedia's reluctance to remove verifiable facts means that a detractor can insert a list of unflattering information, and providing it is well-sourced and neutrally described, it is unlikely to be challenged. The view that negative weight is best countered by positive information being added rather than information being removed can result in bias, as the detractor is not obliged to insert the positive, and no-one else may have the motivation or knowledge.

Such articles are probably not libellous, but are certainly harmful to reputations, and can be highly distressing. Again, Wikipedia's claim to neutrality may lend a credibility to a hatchet job that might be dismissed as biased if appearing on a less-credible website. Also particularly dangerous is the fact that clever, sourced and apparently neutrally worded, biased editing is less likely to fall foul of Wikipedia's checks and balances than simple unsourced untruths. Verdict – highly harmful.

4. Privacy violation[edit]

This is information which is true, but not generally broadcasted. It may include dates of birth, birth names, contact details, school records, family details, and other information not generally available. Often this information is either unverifiable, or only verifiable from primary sources not easily found. Often this information is placed in the article from the personal knowledge of the editor concerned, who may or may not have malicious intent.

Although unverified information is not permitted on Wikipedia, it is not always obvious to patrollers that the information is not of a public nature. Thus this information may remain in the article until the subject complains, by which time it is effectively published. Verdict – highly harmful.

5. Undue broadcasting[edit]

This is where information that is verifiable, but is only published and notable at a local level, is effectively broadcast through Wikipedia. Wikipedia ought to be about recording what is already notable, not about making notable what is otherwise still parochial. To take an example, a story published in a small-town paper with careless fact checking may be ignored by larger newspapers, and rejected by biographers, but it may be hard to eliminate from Wikipedia, as an "according to the small town gazette ..." story is sourced and verifiable. Verdict – somewhat harmful.

Why Wikipedia struggles with BLPs[edit]

In theory, Wikipedia shouldn't have a problem with BLPs. If it maintains its biographies in a neutral and verifiable fashion, then the subject can have no cause for complaint. He doesn't want the truth published? So what?

However, there are certain aspects of the way Wikipedia generates and maintains content that cause it particular problems with biographical subjects.

Eventualism[edit]

Wikipedia is generally eventualist. If an article subject is deemed in principle worthy of an article, Wikipedia prefers to keep the article, in the hope that "eventually" someone will improve it. Bad articles can be currently embarrassing, but Wikipedia excuses them as "works in progress".

This just doesn't work well for biographies. A bad biography now is damaging the subject now. Plus, whilst a neutral editor may "eventually" turn up and fix it, retaining the article is often simply an invitation for the determined non-neutral editor to persist.

BLP patrollers are often frustrated with inclusionists saying that "this needs to be fixed, not deleted", because a) no-one steps up to fix it, b) even if bad material is removed, the patroller is left to watch the article indefinitely, while the keep voter moves on. Whilst people reverting simple vandalism can avoid content disputes, people watching for BLP find themselves continually dragged into "what's wrong with my edits?" debates that wear them down, and tie them up from continuing their patrolling.

Lack of eyes[edit]

Wikipedia works by a theory that articles reach neutrality by a wide number of editors working on them. If someone edits Scientology in a biased way, there are plenty of people ready to revert or improve – and they won't all despise Scientology. Equally it works by the idea that knowledgeable people will look at articles, so that non-apparent falsehoods will be spotted.

However, the less notable the article is, the more prone it is to fewer people being interested, and the more chance there is that the few interested parties will all be biased. Further, if the bias is not immediately obvious, there is less chance with less notable subjects of anyone knowledgeable spotting it.

Many of our biographies are of low-notability subjects, and thus incredibly vulnerable to the motivated biased editor. Further, if the biased or false editing is clever (and particularly if it understands Wiki rules, and includes apparent sourcing), it is unlikely to be spotted on a low-notability biography. At the same time, low-notability biographies are often potentially the most damaging when got wrong, as they may be the only accessible on-line source of biographical information on the subject, and thus opinion-forming in a way that Wikipedia's entry on George W. Bush is unlikely to be.

This is where notorious Wikipedia BLP cases such as articles on Daniel Brandt etc. are bad examples of the problem. When an article becomes that notorious, it will have a host of eyes on it, and a higher chance of bias and untruth being removed.

The less notable the subject:

  1. The less alternative information online. Thus the damage of the bad wiki-bio becomes proportionately bigger.
  2. The less eyes (watchlists) on the article and the greater chance of bad stuff remaining unspotted.
  3. The less chance of Wikipedians knowing enough about the subject to spot less obvious bias or untruths.
  4. The more chance that a POV pusher will be left unchecked.
  5. Once problems are identified, the less chance of anyone caring enough to monitor the article.

Anonymous (and pseudonymous) editing[edit]

The Wikipedia Foundation claims that it is not legally responsible for any libels in articles, but that the user posting the libel is. However, most people posting to Wikipedia hide under pseudonyms or at best identifiable IPs (who are ironically less "anons" than most other users). Now, anonymity does not unnecessarily protect against libel liability, as court orders can, in theory, retrieve information from ISP and WMF.

However, the feeling of immunity that anonymous editing allows can remove one of the basic deterrents against the publication of libellous material on Wikipedia. Anonymity, in this regard, mitigates not only against the subject but can also work against neutral editing. This seems particularly problematic when the reputation of the subject is at stake, but that of his detractor (who may be a notable critic outside Wikipedia), is not.

Obstacles for the harmed subject[edit]

The first problem for the subject is that complaining about an article is "after the fact". No one should be libelled, harassed or abused by Wikipedia. However, if a subject is making a legitimate complaint that has already happened, Wikipedia has an ethical, if not a legal obligation to make sure its powerful web-presence "does no harm" – yet the community tends (and then only grudgingly) to allow methods to remove and mitigate against harm once caused. The community does not accept arguments that would curtail its editorial freedom in order to prevent harm.

The second problem for the subject with a legitimate complaint is that no-one is required to do anything to assist him. The Foundation rejects any legal responsibility, and no-one in a volunteer community is under any obligation to help in any way. The libelled subject is, in effect, begging for charity.

The third problem is that, even if willing, no individual has any power to adjudicate his case and offer mitigation. There is no centralised authority or editorial control. An OTRS op must work within set policies, and even if personally convinced by the subject of the justice of the case, can pretty much be reverted by someone who sees things differently.

A subject unhappy with his biography may pursue two methods to attempt resolution. 1)e-mailing OTRS 2)interacting on wiki.

Difficulty with OTRS[edit]

The primary difficulty with OTRS is often the mismatch between the complainant's reasonable belief that he is corresponding with an "editorial authority" and the fact that OTRS volunteers have little actual editorial control.

Where an article contains obvious libels, or unsourced negative material, it is reasonably simple for the OTRS volunteer to remove the material under the BLP policy. However, the apology offered for the libellous article cannot be accompanied with any promise that the offence will not reoccur.

There is an imbalance between a hard-pressed OTRS volunteer, trying to be helpful, but dealing with dozens of articles per day, and an often highly-motivated agenda-pusher (or multiples) who are likely to return to the scene of the crime hours, days, or months later. If the article is low-notability, or the issues complex, it is almost impossible to interest neutral Wikipedians to become involved, and certainly to sustain involvement to the level that the motivated detractor is likely to sustain.

Where an article is a reasonably written and sourced hatchet job, the problems are greater. First, the OTRS volunteer has to be willing and able to take the time to acquaint himself with the subject, read and assess the sources, and decide whether the article is neutral. This is difficult as

  1. the complainant is often listing numerous alternative sources which the volunteer is invited to consider
  2. it is often difficult to distinguish a complaint about a plausibly written hatchet job from a complaint about a fairly neutral article which is not written the way the subject wants
  3. if the OTRS volunteer intervenes, he will be viewed as a content editor and need to constantly explain his edits and argue the case – this is really asking a lot

Difficulty with wiki involvement[edit]

The other option for the subject is to personally edit the article or otherwise complain on wiki. This poses a number of problems.

  1. Why should he have to do this?
  2. He will be expected to play by the wiki-rules. His detractor may well-know these already, whereas he does not. He is playing at a disadvantage.
  3. The community is far more likely to react negatively to any lack of civility, legal threats, blanking or out-of-process approaches, than to investigate allegations of libel or bias (particularly if these are apparently sourced and well-written). A libelled subject begins by being angry, not knowing the protocol, feeling legally offended, wanting immediately to remove all harmful content, and having no wiki-friends. It is not much wonder he quickly find himself in the role of bad (or banned) user. See WP:DOLT.
  4. Conflict of interest rules play against him.

The Need for permanent vigilance[edit]

Even if the subject manages to obtain a correction, there is no guarantee the correction will persist, since at any future time, any editor (or the same motivated character assassin) can reintroduce false or harmful material. No long term monitoring of the article can be guaranteed by OTRS or the community, and the only way the subject can ensure that the article remains fair is to daily monitor it himself, and then go through the same stressful processes if problems re-occur.

Permanent vigilance may seem an unreasonable burden to place on the subjects of a quarter million BLPs. And is certainly not required if one obtains a correction from any other publication.

Underlying issues[edit]

Lack of responsibility[edit]

a. attitude of entitlement: Wikipedia's users often seem to exhibit a belief in the right to publish anything that isn't specifically illegal or actionable. Wikipedia demands little or no responsibility from those who are given the power of publishing information on perhaps the world's largest reference source. Calls for responsibility and restraint are routinely dismissed as censorship or an infringement of free speech. Whereas most large publications are restrained by journalistic codes of ethics in addition to legal advisers. Matters of taste, decency, fairness, and considerations such as what may displease the reader/customer also limit effective freedom. Wikipedians consistently reject such considerations.

b. lack of community responsibility: The notion that the editor publishing libel is legally responsible for that publication, and the Foundation's rejection of any vicarious liability, often leads Wikipedians to the conclusion that Wikipedia bears no corporate responsibility for harm caused and is under no real moral or legal obligation to engage in harm reduction. Thus, when changes are proposed that might alleviate harm to subjects, the focus immediately falls on the perceived harm to the project, or to the 'wiki' principle of free editing, rather than the potential benefits to BLP victims. The tendency is to think "not our fault, and not our problem". However, if individuals publish libels, it is Wikipedia's structures that allow them to be published, and there is a moral onus on the Wikipedian community to put in place the correct checks and balances to minimise harm.

Privileging of wiki-norms over real-norms[edit]

Breaches of wiki-norms such as civility, consensus and process tend to be of more concern to Wikipedians than issues that affect real people in the real world. The treatment of a justifiably angry subject who after being libelled attempts to amend his biography is a clear example of this. Far more concern is exhibited over the fact that someone is threatening to sue for slander, than the (perhaps obvious) fact that they are being slandered.

Maintainability[edit]

Wikipedia has 818,783 biographies on living people. It never asks "can we maintain these?" or "are we taking all reasonable steps to ensure they are not harming people?" The assumed benefits of allowing anyone to create an article, and anyone to edit it, and keeping articles on thousands of low-notability individuals (even after previous libels have been identified) are assumed to outweigh the cost in terms of damage to the subjects. However, no risk assessment is ever made, because Wikipedians do not bear the risk. In any cost/benefit analysis, the cost are to others, and the benefits to the editors.

"This article can be fixed" syndrome[edit]

Almost any subject can be written about in a broadly neutral and factual manner. Almost any article can be "fixed". However, this obvious fact is often used to effectively negate any mention of the systemic problems. Those concerned about BLPs damaging people are challenged to be specific, and then the specific case is fixed, and the complainer told that no problem remains.

However, the fact remains that we have thousands of unreferenced bios. A percentage of them will be libellous, and a larger percentage grossly unfair. Where is the urgent task force fixing that?

Proposed solutions – assessment[edit]

Any attempts to change Wikipedia's processes to reduce harm to living people needs to be considered against the impact of that change on the wider project. Since neutrality and accuracy are key project goals, any attempt to reduce inaccuracy and bias in biographies should be consistent with that goal. However, Wikipedia has also grown its content through its commitment to the secondary principles of eventualism, inclusionism, lack of centralised control, and that "anyone can edit". Reducing biographical harm will necessarily mean re-negotiating certain aspects of these principles, and that will always be contentious. It must be borne in mind though that elements of those principles are already limited for the greater good of the project (e.g. there are limits to what we include, there are restrictions on what new editors can do – moves, we prevent editing via protection and semi-protection etc).

On the whole Wikipedia has been reluctant to compromise secondary principles to reduce biographical damage. Even the limited rules of the biography policy have been controversial and are still widely disliked. However, Wikipedia has compromised its freedom for the sake of biographical damage in the past. In response to the Seigenthaler incident, Wikipedia removed the ability of non-logged in users to create new articles. Whether that limitation was effective is another judgement to make. Further, that controversial move was made by the authority of Jimbo Wales rather than by community agreement, whether Wikipedia's Godking has the muscle to make such decrees today is questionable, and whether the community could ever reach consensus for such a bold move even more questionable.

Below is a list of many of the proposed "fixes", and some cost benefit analysis of each. What is the cost to the project – what is the benefit to the subject.

The following should be kept in mind when considering any solution.

  1. The desirability of being able to offer a violated subject some reasonable assurance against repeat.
  2. The desirability of increasing the rate at which libels and bias are identified and removed by the community.
  3. The desirability of decreasing the number of under-watched articles (either by increasing the watching, or decreasing the articles)
  4. The desirability of having more effective ways of identifying and curtailing the editor with malicious intent.
  5. The recognition that subject damage is particularly problematic at the lower end of the notability spectrum, where there are fewer informed eyes watching, and Wikipedia is more likely to be a major part of the total accessible on-line information.

Remove all BLPs[edit]

Ease of implementation: Reasonably straight forward, just a new WP:CSD.

Cost to the project: Immense – really, no article on George W. Bush???

Benefit to the subject: High. Problem should largely be gone, although some would transfer to other articles.

Verdict: Not realistic. Although it does raise the question, is there a price for subjects which is simply too high?

Subject's 'opt-out'[edit]

Ease of implementation: Low to moderate. Any subject can e-mail OTRS and on authentication of identity (how?) can demand deletion. OTRS ticket is cited in deletion summary, no re-creation allowed.

Benefit to the subject: Well, no more "Daniel Brandt," "Angela Beesley," or "Don Murphy." It is an easy means of guaranteeing to a violated subject that there will be no repeats. However, the subject still has to know they have a biography and a right not to have one. By the time that happens, a libel or bias may have been published for some time. Further, the subject has to go on public record as declining a wikipedia biography. Perhaps that's a cost most would be willing to pay.

Cost to the project: Debatable. The loss of the Brandts, Beesleys and Murphys is low, however, how many subjects would 'opt out' is unknown. There is the danger of high profile individuals (whose well edited and watched biographies are not particularly problematic) opting out. Or worse, publicists trying to negotiate content changes as the price for "remaining in". What happens if we end up with no articles on Michael Jackson or Tom Cruise?

Verdict: Should be considered, but not as straightforward as it sounds.

Subject's 'opt-out' II (only for lower notables)[edit]

Ease of implementation: Higher. We'd need to define "lower notability" – and probably argue in each case. Unless this is adopted as above, with the exception of subjects who have print biographies or entries in works such as Britannica.

Benefit to the subject: Still, no more "Daniel Brandts" or "Angela Beesleys," but "Don Murphys" would end up in debate. But, the subject may now have to argue on-wiki for their own non-notability. Publicly argue for their privacy. Contend they are not public figures, but do so in a public forum? Could be too big a hurdle.

Cost to the project: We lose the Brandts, Beesleys and (maybe) Murphys but keep Michael Jackson or Tom Cruise. The cost depends where the threshold is set – but the increase in process and argumentation may be great.

Verdict: Certainly worth considering. But falls at the impossibility of trying to define who is "public" or "marginally notable" remember we are talking about a range of people notable for different reasons - no definition is going to be satisfactory. A "dead tree" standard is going to hit against the person who has just recently become VERY famous. (Imagine I shoot George Bush on Monday, and then e-mail OTRS on Tuesday to reject a wikibio.) Most of us would agree Angela Beesely isn't public, but what about block buster film producers. We're simply not going to get a workable definition here. And any case-by-case will just look like afd.


  1. The subject asking I've already outlined the problem with this elsewhere. It is open to people trying to manipulate entries "I'll let you keep my bio, but only if...." and on the other hand it leaves the subject having to ask - if we've got a bio on a not-very-notable subject and it is routinely the target of POV pushing, and hatchet writing, why should we wait until the subject becomes aware of it (when it is already done its damage) before saying "Let's not"

Increase the notability threshold for BLPs[edit]

Ease of implementation:


Cost to the project:


Benefit to the subject:


Verdict:

Semi-protect all BLPs[edit]

Ease of implementation: low, can be done by a bot

Cost to the project: Does this mean no newbie can edit George W. Bush? Seems a high cost.

Benefit to the subject: No more flyby vandalism. But vandalism is the least of it. The determined detractor is not prevented.

Verdict: Tends to focus on "vandalism", which is not really the BLP problem. But it might help. Worth considering, but only just. There could at least be a rule that articles where the subject has complained, or there's been a long term pov pusher, can be permanently semi-protected. However, the fear would be that wikipedians would use it in response to vandalism that annoys the community and doesn't harm the subject "hey Michael Jackson is always having BLP violations posted - he's entitled to semi-protection!!", ignoring the fact that all problems with high-profile articles are quickly seen and reverted. Meanwhile the host of low-notability bios would only get protection after the problem had been spotted (and that may be months after the libel).

Doc's solution[edit]

A solution that does not (necessarily) involve the subject asking, and does not need us to define "marginal notability" - I'd like a way where it is easier to be rid of troublesome BLPs, but the community alone gets to decide on a case by case basis, what to keep and what to remove. Here's my idea:

Change the deletion criteria for AfD to read:

In the case of biographies of living people, where a number of editors have expressed the opinion either
a) that the biography could cause distress to the subject, or
b) that the biography will be particularly difficult to maintain in a fair and accurate state - due to the poor available sourcing or it being of such low interest that few but biased editors will be willing to maintain it,
In such cases, the closing administrator shall close the debate as keep only where there is a consensus that Wikipedia should retain the article. In all other cases the default shall be to delete the article, or to relist where participation has been low.

The advantages are

  1. We get rid of troublesome BLPs
  2. We don't need to involve the subject at all, or consider their views
  3. We don't end up with the nonsense of having bios of four members of a gang, but none for the the fifth, cos he didn't like the idea.
  4. We don't need any new process or policy, we just use afd
  5. We don't have to define "marginal notability" or "private person" - afd can decide each on its merits
  6. No power is given to individual admins or OTRS people - everything gets an open community debate
  7. We don't end up deleting an article because it falls foul of some arbitrary definition - we consider everything on a case-by-case basis. If you can convince people that this BLP is needed, we keep it.
  8. Daniel Brandt, Allison Stokke, Brian Peppers, Angela Beesly articles probably die on the first afd, and don't become community footballs. Even an AfD on Don Murphy is likely to give us a clear answer - although it may be "keep".



Some statistics[edit]

These are not necessarily being claimed as significant. They require interpreting. BLP is taken here as "Number articles in cat:living people"

  • Number of BLPs = 818,783
  • Number of BLPs tagged with {BLPUnsourced} (in Category:All unreferenced BLPs) = 2,990
  • Number of BLPs tagged with {POV} etc =

Is anyone interested in doing some statistically significant number crunching here? Not sure what that would entail.

See also[edit]