Subscribe to

Linden Lab’s ‘Daniel Linden’ recently posted a statement entitled “Keeping Second Life Safe, Together” on the Official Linden Blog. The post encourages Second Life residents to report “broadly offensive” content to Linden Lab, and promises that people promoting or providing such content will be “swiftly met with a variety of sanctions, including termination of accounts, closure of groups, removal of content, and loss of land.”Keeping Second Life Safe, Together

‘Robin Linden’ says the post does not articulate a “new position,” but anyone who has used Second Life for more than a couple of months knows that it represents a significant shift away from Linden Lab’s longstanding hands-off approach to content. As a result, the post has sparked many editorials, protests (some with their own sub-controversies), and discussions.

What exactly is prohibited? “Real-life images, avatar portrayals, and other depiction of sexual or lewd acts involving or appearing to involve children or minors; real-life images, avatar portrayals, and other depictions of sexual violence including rape, real-life images, avatar portrayals, and other depictions of extreme or graphic violence, and other broadly offensive content.”

In other words, “pretty much whatever we feel like.”

I’ll add my voice to the growing chorus. This is a poorly considered, dangerously over-broad, and annoyingly opaque policy statement. It should be reversed for a number of reasons, not the least of which is that it potentially takes Linden Lab out from under the protection of Section 230 of the Communications Decency Act, and thereby exposes it to liability for Second Life users’ creations.

More on that later on. There’s this soapbox sitting here, you see, and one of the privileges of editing VB is that occasionally I get to climb up on it. You can skip to the jump if you want to bypass my rant.

Still with me? First, understand that I’m generally impressed with Linden Lab, and I generally like where they’ve taken Second Life. I’m hardly a regular critic. That said, this is the worst misstep I’ve seen them make, and I think it’s going to take quite a bit of backpedaling to fix it.

The most glaring problem with the “Keeping Second Life Safe, Together” post is that it represents, at best, a half-baked policy statement. For example, the prohibition on “extreme or graphic violence” on its face prohibits every military simulation in Second Life. ‘Robin Linden’ subsequently clarified that “extreme violence implies extreme, disgusting, beyond-the-pale imagery, such as images of violent death, not videogame-style carnage,” but I don’t buy it. Does that mean that if a simulation looks like Hitman it’s okay, but if it looks like Hostel it isn’t? How is watching someone get shot in the face more acceptable than watching someone get slashed with a meat cleaver?

And “depictions of sexual violence?” What does that even mean? Would a showing of Secretary be prohibited? On its face, the policy puts every BDSM-themed area in Second Life in violation. Consider this: depictions of sexual violence are allowed in real life — using people’s actual bodies. The practice is so broadly unoffensive that there is a street fair devoted to it every year in Linden Lab’s own backyard.

But most troubling, in my opinion, is the shiny-happy-people doublespeak that’s being used to defend this move; it’s a new tactic for Linden Lab, which I’d always thought of as refreshingly honest as far as corporations go. Even the title of the post in question is deceptive — what does “safety” have to do with “offensiveness” in the first place? It gets worse. ‘Robin Linden’ defended the move during her office hour by asking, “When faced with an opportunity to create a new world where things are supposed to be better, do you think there’s a place for slavery, forced sex, and the like?” First off, I don’t think that Second Life is a “new world where things are supposed to be better,” I think it’s a three-dimensional communication tool that simulates an environment we’re familiar with, but if you want to play that game, I’ll take the question. You’re damn right there’s a place for it — just like there is in a public library, an art gallery, and a history book.

Your “new world where things are supposed to be better” should not include censorship of the expression of ideas based on what you say your citizens want to see and hear about, even if you do think it’s for our own good. Because even if you are right about this policy this time — and you emphatically are not — the next guy might decide to ban political commentary or art that criticizes religion, and your policy sets the precedent. This is political theory 101 stuff, and if you’re going to act like this is a nation, you better start there.

The absurdity of this question really can’t be overstated — Haley’s Roots could not be enacted in this “new world where things are supposed to be better,” neither could The Color Purple. Should actual slavery and actual forced sex be forbidden? Of course. And they are, by real laws in the real world; Linden Lab runs a simulation. Even if you view it as a nation of some sort — perhaps particularly if you do — the “government” should no more ban depictions of some of these things than a library should ban descriptions of them.

But my voice is just one of many who are saying this policy is a bad idea from a user-relations and political theory standpoint. There are three legal points that are being routinely overlooked in the commentary on this, so I’ll get off my soapbox and get to the law stuff.

First, almost every article I’ve read about this inaccurately accuses Linden Lab of violating Second Life participants’ “free speech rights.” That’s not how it works — the First Amendment of the U.S. Constitution only prohibits government action. If Linden Lab wants to turn Second Life into Club Penguin, it’s allowed to. Like it or not, you don’t have a constitutionally protected “right” to do much of anything in Second Life.

Second, the term “broadly offensive” is very close to the language of a landmark obscenity decision, and that’s probably at least partly why it is being employed here. Although again, we’re not dealing with government action (so Linden Lab is free to apply whatever criteria it likes) and the post from ‘Daniel Linden’ goes beyond sexually explicit material, the “Miller test” for obscenity appears to be instructing the policy. The test comes from the 1973 United States Supreme Court ruling in Miller v. California.

Delivering the opinion of the Court, Chief Justice Warren Burger wrote,

The basic guidelines for the trier of fact must be: (a) whether ‘the average person, applying contemporary community standards‘ would find that the work, taken as a whole, appeals to the prurient interest, (b) whether the work depicts or describes, in a patently offensive way, sexual conduct specifically defined by the applicable state law; and (c) whether the work, taken as a whole, lacks serious literary, artistic, political, or scientific value.

In other words, obscene material is material that is “broadly offensive.” I suppose Burger could have saved a few words if he’d talked to ‘Daniel Linden,’ but at least the Supreme Court ruling includes an exemption for works with serious literary, artistic, political, or scientific value.

Finally, though they say it isn’t so, everyone participating in Second Life recognizes that the ‘Daniel Linden’ post represents a policy shift — and the shift is a terrible idea from a legal perspective. There’s never been a shortage of “broadly offensive” content in Second Life. For better or worse, Linden Lab has always been very hands-off about content. Not any more.

VTOReality beat me to the punch on this, asking, “Are Lindens Heading Into Deeper Legal Water?” The short answer is: “Yes.” I’ll quote VTOReality here, which in turn, quotes intellectual property attorney Michael Bennett:

Up until this very recent ruling providers of interactive computer services have been immune from liability for content created by third parties.

The touchstone of section 230(c) is that providers of interactive computer services are immune from liability for content created by third parties. The immunity applies to a defendant who is the “provider . . . of an interactive computer service” and is being sued “as the publisher or speaker of any information provided by” someone else. 47 U.S.C. § 230(c). “[R]eviewing courts have treated § 230(c) immunity as quite robust.” Carafano v., Inc., 339 F.3d 1119, 1123 (9th Cir. 2003)

But the Fair Housing Council v. suit has take a major twist and turn on this.

Michael Bennett, a partner in the intellectual property department of Chicago law firm Wildman Harrold, said that previous cases confirmed and expanded immunity for ISPs, Web site operators, and listservs.

“Then this one comes down and begins to draw some lines,” he said during an interview Thursday.

The new ruling means that sites that control user-generated content — especially dating and automated brokering sites that use “matching” technology — are considered publishers and are therefore liable, Bennett said.

“If you guide the content too much, or select which of the content will be allowed, you could lose immunity for that portion of the content and be held liable as a publisher,” Bennett said. “The problem arises when nonpublishing Web sites want to put up blogs and forums. Typically, they want to guide the content to some extent, so that it reflects an appropriate image for the site and company.”

Let’s be clear here: is having its feet put to the fire for providing a questionnaire and publishing users comments. That’s relatively passive. Linden Lab, on the other hand, is actively editing user-created content and encouraging users to report “broadly offensive” content for removal. That potentially puts Linden Lab on the hook for any content they miss that violates the Communications Decency Act. Given the size of the user base relative to the administrative team and the simplicity of content creation, there’s just no practical way they’re ever going to be able to get it all.

Hopefully, both for Second Life users and for Linden Lab, someone in legal will take a hard look at this policy and get it reversed or clarified such that it can be reasonably enforced. It is a terrible policy for many reasons, not the least of which is the increase in Linden Lab’s own exposure to liability.

Email This Post Email This Post
Print This Post (Printer Friendly Formatting) Print This Post (Printer Friendly Formatting)

Related Posts on Virtually Blind

13 Responses to “Commentary: Policing Content is a Very Bad Idea”

  1. [...] Benjamin Duranske, intellectual property attorney and virtual law blogger has posted on the shift of Linden Lab’s governance policy, and while Duranske’s post doesn’t contain anything we haven’t said already, he has a certain gravitas in these matters. Time for a roundup. [...]

  2. [...] Benjamin Duranske, editor of Virtually Blind, points out the the term “broadly offensive” seems to be guided by the “Miller test” from the 1973 United States Supreme Court ruling in Miller v. California where the opinion of the Court was written by Chief Justice Warren Burger: The basic guidelines for the trier of fact must be: (a) whether ‘the average person, applying contemporary community standards‘ would find that the work, taken as a whole, appeals to the prurient interest, (b) whether the work depicts or describes, in a patently offensive way, sexual conduct specifically defined by the applicable state law; and (c) whether the work, taken as a whole, lacks serious literary, artistic, political, or scientific value. [...]

  3. on 10 Jun 2007 at 11:05 pmPetey

    I’ve been saying for a while now that content control was a bad idea because of the CDA. Thanks for confirming my suspicions.

    And when do we start a betting pool on when–and why–Linden Lab will be sued to kingdom come. Will it be illegal gambling? Child pornography? Fradulent advertising? (Virtual) property theft (a la Bragg)? Or perhaps even the unthinkable–obscenity?

    Roll the dice, baby.

  4. on 11 Jun 2007 at 7:29 amSarah

    Well put. I never thought that I’d see a policy raise so many red flags. I’ve only taken one class on Internet related law, but one of the central tenants I took away was less control = less liability. I look forward to seeing Linden’s response to the outcry over this decision.

  5. on 11 Jun 2007 at 3:50 pmProkofy Neva

    Re: “shiny-happy-people doublespeak”. The Lindens began years ago encouraging the police-informant state with fostering a climate of narcing out people on the forums. Their inadequate and uneven policing of the TOS on the old (now closed) forums was very much in response to “the community” — which was whatever privileged posse could get started on a witch hunt at any given moment and enjoy biased perception by their old beta friends, residents-turn-Lindens. The very uneven and arbitrary enforcement of the TOS — and blatant lack of enforcement for their friends — set the Lindens up as false arbitrers of the community in any real sense, and set the stage for the “safety” police of May 31.

    The Lindens are also contemplating putting abuse-reporting buttons on the Events list to get rid of all those forms of ads and sales events that certain high-brow cultured folks hate. Watch how the same people who have worked themselves into a frenzy about having freedom of sexual content will become prudes and zealous Calvinists about entrepreneurial content on the Events list, frothing at the mouth to have it removed, and stamping out low-brow commerce in the process. So very SL!

    I’m not finding the invocation of Fair Housing Council v. at all persuasive here and I think this needs to be challenged, whatever the fears about it.

    The Lindens haven’t edited anything or censored anything explicitly sexual — yet. They’ve removed 108 ads from the search mainly having to do with casino advertising, which they banned a few weeks ago, and a few explicit ageplay art sites, which they also indicated was prohibited. They’ve banned the two German residents alleged to have also possessed RL child porn.

    As for the issue of what’s “broadly offensive,” in fact, the Lindens are resisting interpretations of this, leaving it overbroad in fact (and that’s what some people find a problem) and emphatically stating that it’s “the community” that has to make the determination of what is broadly offensive.

    Where’s the guidance, then?

    If the Lindens are not editing, are not unilaterally flying around removing content, but are responding to abuse reports and saying “the community will decide,” we can complain that this is fake (I sure do), but we can’t say it’s like Fair Housing Council precisely because they are NOT guiding and NOT editing but telling us once again, “Your world, your imagination” and urging *us* to do ths guiding.

    It puts the onus squarely back on the user — he will answer within his own jurisdiction, should that RL jurisdiction’s authorities decide to act. It no longer makes LL available as “cover” for people’s autonomous zones.

    I think if you are going to invoke Fair Housing Council, you need a thorough argumentation illustrating how Linden “guides”. And for that we need some jurisprudence. They haven’t acted yet. There’s no BDSM group closed. There’s no really gross snuff site deleted. There isn’t a single action.

    And as I maintain, what people have to do in this uncertain and potentially repressive and arbitrary environment is not file abuse reports so as not to become part of the illegitimate force defining what’s “broadly offensive” when the community hasn’t really been polled or involved.

  6. on 12 Jun 2007 at 3:29 amRob Danton

    I agree with Prokofy on this… LL has always seemed to have a “the community will police itself” approach. Philip Rosedale has been quoted many times as saying he doesn not want to get involved in law enforcement. probably for the reason that to start enforcing standards makes it liable.

    So this latest statement is a fudge, a vague sop to those who, thanks to the latest ageplay related furore, are calling for a clampdown. LL wants to be seen to be doing something. HoweverI can’t see that the latest post really makes for much of a break from their old standpoint – LL is not suggesting that it will police content any more than before. It is asking that residents report “broadly offensive” stuff.

    I can see why LL had to act on the casino thing. The last thing the company needs is to take on the US Government right now. Its not in our or its interest to do that. The same goes if these German residents had RL child porn. It would be impossible for LL to defend the rights of paedophiles.

    We are all adults (we are, right?) and as Prokofy says the onus is on us as residents to continue to live and let live. I’ve seen stuff in Gorean sims that I think is deeply offensive to women in particular… but am I going to report it? No. I’ve gone to that sim to see what it’s all about – and it seems horrible to me so I won’t go back.

  7. on 12 Jun 2007 at 8:40 amBenjamin Duranske

    ‘Prokofy’ — thanks for the comments. One point made here is definitely worth amplifying: this isn’t settled law, it’s a new theory that is only being mapped by the courts on a case by case basis.

    That said, I still think they’re on thin ice. Here’s at least one specific activity that I think LL can get into trouble for under the ruling: editing people’s “places” descriptions to remove content. So far, it appears to be casino ads and the term “lolita” that were targeted. If you search for the phrase “description removed,” in “places,” in-world, you can see around 100 instances of this editing.

    As ‘Prokofy’ suggests, I think it’s a good idea to look more carefully at the ruling. Here’s what I see there: the Court draws a fairly bright line between, on one hand, simply publishing what users put up (even with minor editing) and on the other hand, participating in the creation process. It finds that participating in the creation process is what triggers the loss of CDA immuity under Section 230(c).

    The most interesting language comes from a footnote. There, the Court writes:

    “When users select the option “I will not live with children,” Roommate publishes this response as “no children please.” The Councils argue that this alteration makes Roommate a content provider and therefore not immune under the CDA for publishing this statement. However, minor editing that does not affect meaning is protected under the CDA as the “usual prerogative of publishers.” See Batzel, 333 F.3d at 1031 (publisher of a newsletter protected under the CDA for making minor editing changes to user-submitted content). Because “no children please” is materially the same as “I will not live with children,” Roommate does not lose its CDA immunity because of the change in wording. On the other hand, the Councils allege that Roommate takes members’ blank selection in the children field and publishes it as “no children please.” We could not find support for this proposition in the record, but if Councils’ allegation is true, then Roommate significantly alters the meaning of the information provided by its members and is not entitled to CDA immunity for posting the resulting content.”

    Given this language, if I were LL’s lawyer, I’d tell them to be careful searching for and removing content — particularly content that might violate the CDA, like “lolita” links. It seems to me that under this ruling, they’re at least potentially on the hook for any they miss.

    Rob Danton – Thanks for the comment. Its interesting to me that you don’t see this as a break from previous policy. Rightly or not, most residents seem to think it is. I wonder how much of that is because residents are viewing it in conjunction with a number of other recent activities alluded to above. The big difference I see is what I just pointed out in my response to Prokofy’s comment: they’re editing people’s content (see the “description removed” discussion above) and whether that’s officially part of this statement or a seperate action that goes along with it, it is definitely a relatively new step for them to be taking, and I believe it’s what can get them in trouble with the CDA.

  8. on 19 Jun 2007 at 10:01 pmYourWorldBlahBlahBlah

    I no longer support the success of the Second Life platform.

    In related news:

    Riyadh, 2010 (Reuters), As active Iranian users continues to climb, Linden Lab CEO Philip Rosedale tells Mitch Wagner of Information Week, “We have never allowed avatars to have virtual sex with other avatars.”

    I’d ask for the time I wasted of my life back, but I already know the answer is “report resolved.”

    Have fun.

  9. on 22 Jun 2007 at 12:05 amTim

    That was an excellent article. Thanks.

    If we were to assume that SL new policy was enforcable and they had an effective ban on all the things they stated, would they go bankrupt?

    The ‘adult’ side of SL is huge. It must contribute to a very sizable portion of the SL economy, maybe even most of it. Thinking of the clothes, bodies, props, maybe they wouldn’t get banned, but there wouldn’t be so many people buying them.

  10. on 22 Jun 2007 at 12:47 amProkofy Neva

    I find the scare quotes around avatars’ names to be a real affectation. It’s not done in journalism, in fact. In fact, we can see what journalistic practice has been: to name the avatar’s name, and identify their RL name. I speak as someone who, in the RL media, has provided my RL name. But this blog is just a blog. It’s not a scholarly journal or a law office’s publication. So it seems contrived to be fussing with scare quotes.

    I stated that 108 pages of content were reviewed and text removed, so I’m not sure why that was reiterated as if it were new news.

    Reading the footnote, I don’t see how removing “Lolita” in keeping with a ban on “ageplay” content is “participating in the creation process”; it’s the opposite of the example given, as content is removed, not blank lines filled in with content.

    I would have to hear what precedents there are for interpreting removal of a sign in violation as being the same thing as “editing content”.

    The intent of this lawsuit — to try to affirm fair housing against a service that evidently some found as being discriminatory — is a different kind of intent than what is found around the SL issues. And the nature of SL itself isn’t as clear as what is.

    I don’t think the CDA will bother with them at this juncture.

    What we have yet to see is any slew of abuse reports and actions. I specifically asked Daniel Linden about this, and he turned the question into a joke, saying people abuse-reported *him* for the policy.

  11. on 22 Jun 2007 at 2:07 amBenjamin Duranske

    The policy regarding avatars names is permanently linked on the front page, in the upper right corner. In sum, I don’t identify people by their real names unless they want to be identified that way. I enclose avatar names in single quotation marks when I don’t have permission to use a real name or when I don’t have a real name. Regardless of the gender of the person behind the avatar, I use pronouns that match the gender of the avatar if I do not have permission to use a real identity.

    I think Prokofy is actually right here though — the policy is important in articles, but in comments, it does seem like an affectation. I think that’s because it is generally understood that most  people leave comments under pseudonyms. Anyway, going forward, when addressing a commenter either directly or by reference, I’m comfortable breaking this guideline (as I did in this very paragraph, and almost certainly have before).

    For those that are interested, the reason for this policy in articles is that SL’s naming convention makes it hard to tell when someone is using a real identity and when they are only willing to be identified as an avatar. Mainstream journalistic standards say that you shouldn’t quote anyone who won’t reveal his or her real identity to you (you can, of course, keep it secret, but you are generally supposed to know who you are quoting). Most — if not all — Second Life bloggers ignore this completely and interchange avatar names and real names in posts without comment or annotation. That’s a reasonable choice, but it’s not my choice. Instead, I chose the single-quote compromise because though I support the theory behind the journalistic standard, I quickly realized it’s nearly impossible to cover virtual spaces, particularly legal issues in virtual spaces, without at least occasionally quoting people who choose to only be identified by their avatar names.

    To that end, I generally ask story sources and subjects for permission to use their real names. When the real name is unavailable or the person is unwilling to allow me to use it, I use single quotes to offset the avatar name. I think single quotes are less “scary” scare quotes than double quotes (which carry a heavier implication of skepticism in my mind) but they do serve to remind the reader that I do not know, or do not have permission to provide, the real name behind the avatar.

    I expect 3-5 years from now a standard will develop for this because the mainstream press is going to have the same problem, but until then, this is my solution. It may not be the best solution, and I’m certainly open to other suggestions that accomplish the same thing.

    Sorry for the long-winded explanation folks, but I’ve put a fair bit of thought into this. This is the public face of my law practice (I’m actually not practicing at the moment as I’m working on a novel, but I plan to re-enter practice in the next year or so, very likely with an emphasis on virtual law, and this blog is partly to let me keep a professional finger on the pulse of this area while I’m not practicing). As a result, I approach VB with somewhat more emphasis on professionalism and journalistic standards than typical blogs exhibit. If some of the decisions that stem off of that seem like affectations, well, I apologize, readers, but that’s the way it’s going to be.

    But like I said, in comments, it’s pretty much understood that most people comment under pseudonyms, so I’ll drop the quotes. Good advice, Prokofy.

    Regarding the legal impact of the ruling, sure, it’s unsettled. Probably will be for some time. No real disagreement there, though that hardly means its not important for a corporation to pay attention to it. Most of IP law is in a state of flux to some degree. Moreover, in this specific case, changing content from whatever it used to be to “content removed” looks an awful lot like editing to me (definitions of “edit” include “To eliminate; delete” and ” To modify or adapt so as to make suitable or acceptable.”) But whatever the additional risk, and it seems inescapable that there is at least some, why take it on? That’s not entirely a rhetorical question, though I think it’s one Linden Lab failed to consider sufficiently. One possible explanation is that they’ve done the calculus and decided that it’s safer to comply with foreign laws even if they’re exposing themselves to increased liability under the CDA, but I haven’t seen anything that makes me think this was that well thought out. To me, it feels more like a stumble than a step, and from a legal standpoint, it looks like a stumble in the wrong direction. In any case, I can’t agree that it’s unlikely the CDA will get used here. It’s just too perfect an opportunity for a zealous prosecutor who is looking for a headline given all the press SL has received and the scope of material that’s available.

  12. [...] Readers with good memories will recall that about three months ago, VB ran a short piece on an odd item from the international desk. China planned to impose a new law which would require that online computer games generate half-points after three hours of play, and zero points after five hours. Games that did not comply would be shut down. Three months ago I made fun of it. Now, I’m worried about what it might mean for Second Life. It’s amazing what a child porn allegation from Germany and an ill-considered policy statement from Linden Lab can do to change one’s attitude. [...]

  13. on 24 Apr 2008 at 12:51 amA key moment for Metabrands? : Kzero

    [...] Commentary: Policing Content is a Very Bad Idea [...]

Leave a Reply

Notes on Comments: Your first comment must be manually approved, but after it is you'll be able to post freely with the same name and email. You can use some HTML (<a> <b> <i> <blockquote> etc.) but know that VB's spam blocker holds posts with five or more <a> links. VB supports gravatars. Got a gravatar? Use the associated email and it'll show with your comment. Need one? Set it up for free here.