User-driven websites — also known as online intermediaries — frequently come under fire for disabling user content due to bogus or illegitimate takedown notices. Facebook is at the center of the latest controversy involving a bogus takedown notice. On Thursday morning, the social networking site disabled Ars Technica’s page after receiving a DMCA takedown notice alleging the page contained copyright infringing material. While details about the claim remain unclear, given that Facebook restored Ars’s page yesterday evening, it’s a safe bet that the takedown notice was without merit.

Understandably, Ars Technica wasn’t exactly pleased that its Facebook page — one of its top sources of incoming traffic — was shut down for seemingly no good reason. Ars was particularly disappointed by how Facebook handled the situation. In an article posted yesterday (and updated throughout the day), Ars co-founder Ken Fisher and senior editor Jacqui Cheng chronicled their struggle in getting Facebook to simply discuss the situation with them and allow Ars to respond to the takedown notice.

Facebook took hours to respond to Ars’s initial inquiry, and didn’t provide a copy of takedown notice until the following day. Several other major tech websites, including ReadWriteWeb and TheNextWeb, also covered the issue, noting that Ars Technica is the latest in a series of websites to have suffered from their Facebook page being wrongly disabled. In a follow-up article posted today, Ars elaborated on what happened and offered some tips to Facebook on how it could have better handled the situation.

It’s totally fair to criticize how Facebook deals with content takedown requests. Ars is right that the company could certainly do a much better job of handling the process, and Facebook will hopefully re-evaluate its procedures in light of this widely publicized snafu. In calling out Facebook’s flawed approach to dealing with takedown requests, however, Ars Technica doesn’t do justice to the larger, more fundamental problem of bogus takedown notices.

As Mike Masnick explains on Techdirt, U.S. federal laws strongly discourage online intermediaries from trying to figure out if takedown notices are legitimate or not. If Facebook were to refuse to comply with a copyright takedown notice that subsequently turned out to be meritorious, it would lose its safe harbor provided for in 17 U.S.C. § 512(c). Should Facebook err in its judgment, therefore, it would potentially be on the hook for harsh copyright infringement penalties. In effect, the DMCA incentivizes what Masnick describes as “massive overreactions” by online intermediaries.

That’s not to say that there aren’t some simple steps Facebook could take to combat bogus takedown notices without exposing itself to additional liability, especially in “easy” cases, as Ars and others have argued. Verifying that takedown notices are associated with valid email addresses is one such step that Facebook apparently does not currently employ. Facebook could also be more responsive to users whose content has been disabled, at least when the content in question is highly visible.

Perhaps more importantly, Facebook should adopt a system for enabling users who believe their content has been wrongly disabled to file a counter notification. YouTube, for instance, has a slick online system that lets users challenge wrongful takedown requests. Under 17 U.S.C. § 512(g), an online service provider may restore previously-disabled content between 10 and 14 days after receipt of a valid counter notification if the content owner hasn’t initiated legal proceedings. It’s odd that Facebook hasn’t adopted an online counter notification system, especially given that service providers are shielded from liability if they respond to counter notices in accordance with section 512(g).

While it would be great if Facebook were to manually and thoroughly screen all user complaints and requests, expecting online intermediaries to pay for a live human being — say, an intellectual property lawyer or a paralegal — to vet the legal merits of each takedown notice is simply unreasonable. Facebook has more than 600 million active users, but a mere 2,000 or so employees (although that number may soon grow substantially). That’s over 300,000 users per employee!

And let’s not forget that Facebook is a free service. The company generated a scant $4 of revenue per user in 2010. Facebook’s going to have to do a much better job of monetizing its platform before we can reasonably expect it to vet legal requests on its users’ behalf. Even Google — with a head count and revenue more than ten times Facebook’s — is frequently chastised for not doing enough to identify bogus or otherwise invalid takedown notices. Based on some of the “horror stories” that have been reported recently, Ars Technica is lucky that Facebook restored its page within a day of its removal.

Even if Facebook improves its system, however, the underlying problem of bogus takedown notices is probably here to stay — that is, until Congress acts. Reopening the legislative debate over the DMCA is a risky gambit, but at least in theory, Congress could improve the statute by adopting some relatively minor tweaks.

First, the DMCA should do more to deter parties from filing invalid or bad faith DMCA takedown notices. Courts rarely punish parties for filing illegitimate takedown notices, as it is very difficult in practice to show  that a notice was filed in bad faith. All in all, the overwhelming majority of incidents of bogus takedown notices go unpunished, as I’ve discussed before on these pages.

Wendy Seltzer of Princeton’s Center for Information Technology Policy chronicled the chilling effects of DMCA takedown abuses in a recent Harvard Journal of Law & Technology article. She suggests a few legislative fixes to 17 U.S.C. § 512(f) to better balance the interests of users and rightsholders:

The law should require greater diligence: declarations on penalty of perjury to match those required by the respondent, and perhaps even a bond against erroneous claims. . . . Strengthening the counter-suit provisions could encourage a plaintiffs’ bar to take up these cases as private attorneys general. Stiffening the penalties against claimants who obtained takedowns through misrepresentation of infringement would encourage claimants to verify and support their claims of infringement or penalize them for failure to do so rather than allowing them to shift that burden to service providers and posters.

Congress should also create a safe harbor, notice-and-takedown system for online trademark infringement, as Elizabeth Levin has argued. While copyright takedown notices receive most of the attention in the IP debates, there’s no DMCA-esque process established in statute to provide for online intermediaries to disable and repost allegedly trademark-infringing content.

Doug Feaver, a former Washington Post reporter and editor, has published a very interesting editorial today entitled “Listening to the Dot-Commenters.”  In the piece, Feaver discusses his personal change of heart about “the anonymous, unmoderated, often appallingly inaccurate, sometimes profane, frequently off point and occasionally racist reader comments that washingtonpost.com allows to be published at the end of articles and blogs.” When he worked at the Post, he fought to keep anonymous and unmoderated comments off the WP.com site entirely because it was too difficult to pre-screen them all and “the bigger problem with The Post’s comment policy, many in the newsroom have told me, is that the comments are anonymous. Anonymity is what gives cover to racists, sexists and others to say inappropriate things without having to say who they are.”

But Feaver now believes those anonymous, unmoderated comment have value because:

I believe that it is useful to be reminded bluntly that the dark forces are out there and that it is too easy to forget that truth by imposing rules that obscure it.  As Oscar Wilde wrote in a different context, “Man is least in himself when he talks in his own person. Give him a mask, and he will tell you the truth.”   Too many of us like to think that we have made great progress in human relations and that little remains to be done. Unmoderated comments provide an antidote to such ridiculous conclusions. It’s not like the rest of us don’t know those words and hear them occasionally, depending on where we choose to tread, but most of us don’t want to have to confront them.

It seems a bit depressing that the best argument in favor of allowing unmoderated, anonymous comments is that it allows us to see the dark underbelly of mankind, but the good news, Feaver points out, is that:

But I am heartened by the fact that such comments do not go unchallenged by readers. In fact, comment strings are often self-correcting and provide informative exchanges. If somebody says something ridiculous, somebody else will challenge it. And there is wit.

He goes on to provide some good examples.  And he also notes how unmoderated comments let readers provide their heartfelt views on the substance of sensitive issues and let journalists and editorialists know how they feel about what is being reported or how it is being reported. “We journalists need to pay attention to what our readers say, even if we don’t like it,” he argues. “There are things to learn.”

I applaud Mr. Feaver for this.  This is a struggle not just for journalists at major media outlets but also for bloggers like us here at the TLF.  There are times when very annoying, even hurtful things are said by anonymous commenters here at the TLF. Our policy, however, has generally been to allow a vibrant exchange of views, except in the rare circumstances where the commenter utters racial epithets or starts issuing death threats. Or, if a specific commenter goes into “stalker mode” and does nothing but post harassing, irrelevant comments all day, then those will occasionally be discarded. But, generally speaking, it’s “anything goes” here. (We even allow spam!)  Each author, however, is free to decide for themselves where to draw the line, but we all generally err on the side of completely unmoderated exchange for the reasons Feaver lists.  We know it is far more likely that we’ll get hostile anonymous comments rather than nice ones, but it’s good to get feedback of all varieties, even when it’s nasty.

From a policy perspective, however, this issue is taking on greater weight because some folks believe that unmoderated, anonymous user comments result in harassment, hate speech, defamation, or privacy violations.  As a result, there has been a growing chorus of critics who claim something must be done to remedy this problem.  Cass Sunstein and Richard Thaler, for example, have advocated a Civility Check that “can accurately tell whether the email you’re about to send is angry and caution you, “warning: this appears to be an uncivil email. do you really and truly want to send it?”” The state of Kentucky has considered legislation that would ban online anonymity, even though that would be clearly unconstitutional. Respected law school professors such as Mark Lemley and Daniel Solove have toyed with the idea of DMCA-like “notice-and-takedown” regime for potentially defamatory comments online.  I once even heard Cal-Western law school professor Nancy S. Kim suggest that blogs, social networking sites, and other interactive sites should institute a “cooling off period” to address cyber-harassment. By requiring all those seeking to comment to wait a certain length of time before a message or image is posted to a website, she hoped that some commenters might choose to tone down or even remove the potentially offending messages or images.

Of course, it is more likely that readers would just choose not to comment at all if any of these proposals where enshrined into law.  And that gets to the heart of what’s wrong with any potential legal response to this “problem” of online anonymous, unmoderated speech and user comments:  It will massively chill free speech and expression.  Sure, that would get rid of the hecklers and the jackasses who cause grief for some, but it would also deprive us of the many constructive user comments and criticisms that make the online experience — for better or worse — the most open, vibrant exchange of views ever known to man.

Finally, it goes without saying that this debate is fundamentally tied up with the future of Section 230 and the question of intermediary liability.  Currently, online service providers of all flavors are generally not required to police or screen user comments or force users to be authenticated and reveal their identities before posting comments.  Section. 230 has been the key to protecting intermediaries from punishing liability that would otherwise force them to severely curtail online expression, or run the risk of being driven under by the weight of endless lawsuits.  This is why I have argued that Sec. 230  is “the cornerstone of ‘Internet freedom’ in its truest and best sense of the term.”  But, again, all this could change if we are not vigilant in defending Sec. 230.

OK, now that I’ve made this impassioned defense of unmoderated and completly anonymous online exchange, let the hateful comments fly!

David Margolick has penned a lengthy piece for Portfolio.com about the AutoAdmit case, which has important ramifications for the future of Section 230 and online speech in general. Very brief background: AutoAdmit is a discussion board for students looking to enter, or just discuss, law schools. Some threads on the site have included ugly — insanely ugly — insults about some women.  A couple of those women sued to reveal the identities of their attackers and hold them liable for supposedly wronging them.  The case has been slowly moving through the courts ever since. Again, read Margolick’s article for all the details.  The important point here is that the women could not sue AutoAdmit directly for defamation or harassment because Section 230 of the Communications Decency Act of 1996 immunizes websites from liability for the actions of their users.  Consequently, those looking to sue must go after the actual individuals behind the comments which (supposedly) caused the harm in question.

I am big defender of Section 230 and have argued that it has been the cornerstone of Internet freedom. Keeping online intermediaries free from burdensome policing requirements and liability threats has created the vibrant marketplace of expression and commerce that we enjoy today. If not for Sec. 230, we would likely live in a very different world today.

Sec. 230 has come under attack, however, from those who believe online intermediaries should “do more” to address various concerns, including cyber-bullying, defamation, or other problems.  For those of us who believe passionately in the importance of Sec. 230, the better approach is to preserve immunity for intermediaries and instead encourage more voluntary policing and self-regulation by intermediaries, increased public pressure on those sites that turn a blind eye to such behavior to encourage them to change their ways, more efforts to establish “community policing” by users such that they can report or counter abusive language, and so on.

Of course, those efforts will never be fool proof and a handful of bad apples will still be able to cause a lot of grief for some users on certain discussion boards, blogs, and so on.  In those extreme cases where legal action is necessary, it would be optimal if every effort was exhausted to go after the actual end-user who is causing the problem before tossing Sec. 230 and current online immunity norms to the wind in an effort to force the intermediaries to police speech.  After all, how do the intermediaries know what is defamatory?  Why should they be forced to sit in judgment of such things?  If, under threat of lawsuit, they are petitioned by countless users to remove content or comments that those individuals find objectionable, the result will be a massive chilling effect on online free speech since those intermediaries would likely play is safe most of the time and just take everything down.

Which brings up back to the danger of a 230 backlash following the AutoAdmit case. As Margolick notes of the case:

By any standard, the plaintiffs’ catch has been meager. Even with one of the country’s top intellectual-property lawyers, backed by a super-elite law firm, going after them, most of the worst offenders got off scot-free. The fact that so few prey were netted could prompt calls to modify Section 230(c), if only to give victims of internet abuse more of a chance. Brian Leiter, the professor and vocal critic of AutoAdmit, sees it coming. He calls the free pass enjoyed by Google and other carriers “a disaster” and says change is inevitable. “The point at which some senator’s daughter becomes the target of this kind of campaign of online vilification and harassment on the next iteration of AutoAdmit — something’s going to happen,” predicted Leiter, who now teaches at the University of Chicago Law School.

Unfortunately, although I obviously don’t agree with Prof. Leiter about Sec. 230 being “a disaster,” I think he’s right to assume that one particularly visible and sensitive case could end up bringing 230 back up for political reconsideration.  Commenting on this on the Info/Law blog, William McGeveran of the University of Minnesota Law School summarizes my own feelings regarding the potential danger we face going forward:

the response will be a dramatic evisceration or even elimination of Section 230 immunity. We might end up with some kind of notice-and-takedown regime that could be abused just as it is in the DMCA setting, allowing anyone to effectively force the elimination of web content they dislike with the mere untested allegation that it was tortious. Worse, we might see an effort to repeal section 230 altogether, making it impossible to run an open online forum for user-generated content without risking significant liability.

Indeed, there have already been calls for variants of a notice-and-takedown regime for speech put forward by law professors such as Mark Lemley and Daniel Solove. And the rising calls at the state level for legislation to address cyber-bullying could become another pressure point in the movement to deputize the middleman.

Importantly, however, the “[AutoAdmit] case has already made a difference,” Margolick notes:

Things have calmed down on AutoAdmit, where, Cohen says, he’s driven away the worst actors and enlisted volunteer moderators. Some post­ers, moreover, have announced their “retirement”; any further self-expression, they’ve concluded, is clearly not worth the risk. Thanks to the case, casual defamers — those who take potshots for sport — may now refrain out of empathy for the plaintiffs, while the more malicious may have been intimidated into silence. The case may also have helped Heller and Iravani [the plantiffs in the case] cleanse their Google pages, as the old slurs have fallen farther down the screen. And last spring, Cohen quietly removed the offending threads. He’d have done so sooner, he says, had he been asked more nicely.

This gets back to my point about how self-regulation, social norms, and public pressure can be an effective way to counter online harassment without resorting to major changes in law or liability norms. Again, I think Sec. 230 is worth preserving and the efforts to tinker with it are likely to open a Pandora’s Box of problems for intermediaries and average users alike.  The question going forward is, do we let the presence of a few bad apples online justify the complete upending of a legal standard that has made the Internet the most vibrant platform for free speech that the world has ever known?  I certainly hope not.

Some additional reading on the case:

• Eugene Volokh’s thoughts on the case (make sure to read the comments too)
• Eric Goldman’s thoughts on the case
• Amir Efrati of the Wall Street Journal’s Law Blog
• Dave Hoffman at Concurring Opinions Blog (make sure to read the comments too)