Student Blogger - Chicago’s Best Ideas: Dean Saul Levmore on “The Internet’s Anonymity Problem”

Update: Dean Levmore's talk is now available as an .mp3 file.

The walls of the Law School's bathroom stalls used to display the student body's complaints about professors and fellow students, but the internet made those walls obsolete. Is the internet different from the bathroom stalls in some fundamental way? Does the internet mark a break from the paradigm of previous media? Dean Levmore does not think so.

On Tuesday, November 11, Dean Saul Levmore gave a talk on "The Internet's Anonymity Problem" as part of the Chicago's Best Ideas lecture series. His main contention is that the internet is not different from other media and should be subject to the same legal regime. Currently, it is not; § 230 of the Communications Decency Act provides that internet service providers (ISPs) are not publishers with regard to user-generated content, so they are for the most part not responsible for online torts committed by their users. (One questioner pointed to Fair Housing Council of San Fernando Valley v Roommates.com (9th Cir 2008) (en banc), where Judge Kozinski wrote an opinion holding a website liable under the Fair Housing Act for discrimination committed by its users, but Levmore remarked that the case is an outlier because, after all, it was written by Judge Kozinski.) If a newspaper, on the other hand, publishes a defamatory letter to the editor, the newspaper may be sued. The most commonly cited reason for the nonpublisher rule in the legislative history of the Act is that the internet is a new medium, so it should be allowed to develop and flourish. But the Act was passed twelve years ago, and the internet has matured since then, so it is time to take stock.

Levmore illustrated the problem with the website Juicy Campus, which I will not link to; obviously, many similar websites exist. The website categorizes gossip message boards by school: one for UCLA, one for Tennessee Tech, one for University of Chicago (which is rather sparse), and so on. The homepage advertises—right in the middle—that posting is completely anonymous, no registration required. Users may post whatever they like, and they do. Levmore gave an example of a typical message thread: Someone asks, "Who is the biggest sorority slut on campus?" Someone responds, "I saw Amy X doing such-and-such behind Biddle Hall." Seven people give props for the post and ask for more detail. Another person gives a modest defense and says that Amy X isn't so bad, to which a poster replies, "Not only is Amy X a slut, she's fat." Someone asks if anyone can think of someone fatter. Several posters throw out some names. And so forth.

Thanks to the different legal regime, the internet does not exert the same controls on speakers as other media do. Television, for example, is subject to a fair amount of government regulation, and consumer demand constrains what appears on television as well. (Consumer demand is the real reason why Big Bird will not drop the f-bomb on Sesame Street any time soon.) Bathroom stalls are cleaned occasionally, and the harm is less because--unlike Juicy Campus--they are not searchable. Soapboxes, such as in London's Hyde Park, are regulated by social sanctions; everyone can see who the speaker is. Newspapers are occasionally subjected to defamation lawsuits, but they are more often protected from speech regulations. Consumer demand prevents them from publishing anonymous, defamatory letters to the editor. But the internet does not have any of those controls. Posting is anonymous, and ISPs usually refuse to give up the names associated with IP addresses. Thus, neither the ISP nor the speaker face legal liability, and the speaker is shielded from social constraints. Consumer demand does not help; a website can support itself with only niche demand, but any internet user may stumble across the page because the message boards are searchable.

Levmore recommends eliminating the liability restrictions on ISPs and forcing them to divulge the identity of IP addresses if subpoenaed. The logic underlying his recommendation is a hypothetical bargain among all users of the internet. The benefit to each person is tiny from being able to post degrading insults about others, but the cost of being a target of these insults is very high. Even if the chance of being targeted is small, the cost is large enough that the expected value outweighs the miniscule benefit. Thus, the bargainers would not immunize such conduct. Levmore focused on differences in expected costs and benefits to each person while implicitly assuming homogeneous preferences, but an alternative formulation could depend on differences in preferences. No one wants to be targeted, but only a minority wants to target others, so the majority demands the minority give up its antisocial behavior.

The present regime presents a conundrum: politicians tend to overreact to problems, yet here the reverse happened. Gangs are an example of the general principle. We do not really know what proportion of violent crime is caused by gangs, but politicians provide a wide array of legislation targeting the gang problem. Yet with the internet, the problem resulted in additional protection for the perpetrators. Levmore provided the classic public choice explanation. The people subject to the regulation are a small, passionate, organized group; the potential victims are the public at large, so their interests are diffuse. Organized groups typically trump diffuse groups when collecting government payouts.

A possible addition to that explanation for exceptional rules for the internet is gender disparity. The victims are disproportionately female, and the unfettered internet's most vocal defenders are mostly male. Not only are the victims mostly female, the hurtfulness of many of the comments is premised on the target's gender. Levmore implicitly acknowledged these facts through his use of examples (though he doesn't discuss it): his paradigm example was "Amy X is a slut," and "fat" and "small-breasted" were other examples. He does not press further because the vast majority of examples would be too vile for the classroom. (This is not to say all remarks are directed towards women; sometimes speakers target businesses as "cheats," and sometimes speakers target males such as, for example, Levmore himself.) Those who want the internet to be completely unregulated, on the other hand, are much more likely to be male. As a rough proxy, more than 80 percent of undergraduate engineering students are male. Those with a more technical bent are keenly aware of the benefits of a free flow of information on the internet, but--since this group is mostly male--they will undervalue the costs of antisocial behavior on the internet. Disparaging comments are less commonly directed at them, and the most harmful comments lose their meaning if directed at males. A skewed understanding of the costs and benefits translates into the policy choices that the organized internet interest group lobbies for at the expense of the diffuse potential victim group.

This additional explanation points to a gap in Levmore's framework: some theory explaining why posters post. Levmore takes the economist's approach of treating preferences as a given; essentially, he gets by on the assumption that if you build it, they will come. A more sociological explanation of why speakers have the preferences they do could be illuminating, and it may have policy implications. Traditional sanctions are inadequate if the motivation underlying disparaging speech differs on the internet than in traditional media; however, if the motivation is the same regardless of medium, then traditional sanctions should fit well. The basic outline of the behavior is the same in both cases, but the cost of engaging in it is significantly lower with the internet. Adding marginal members because of reduced cost may or may not change the group's characteristics. More investigation is needed on this point.

To end, Levmore gave a prediction for the future: internet entrepreneurs will eliminate user-generated content or require registration to post. To attract consumer attention, the value of the information a website provides must be high. Since anonymous, insulting posts have zero value, the average value of the information on a website decreases when it permits those posts. Some websites are apparently already doing this. Wikipedia has an army of 75,000 registered editors to keep its entries in line. Amazon and CNet use robots to track swear words and other mechanisms to keep their customer reviews (relatively) high value. YouTube has always prohibited pornography and has recently become much better at keeping copyrighted content off its site. The University of Chicago Faculty Blog moderates its comment boards. Levmore provided the grounds to believe these developments are positive.

Comments

[Michael F. Martin]

Levmore's idea of treating Internet speech like other forms of speech is like the proverbial common sense that is, in fact, all too uncommon.

[Michael F. Martin]

Also, Chicago should be proud of that the Chicago School solution to this particular tragedy of the commons has been adopted so widely. As if it weren't enough to have one of its faculty President!

[Nathan Richardson]

http://www.juicycampus.com/posts/gossips/all-campuses/

[Luke Gilman]

Very interesting talk. It occurs to me, however, that insults are not the only use of anonymous writing. Anonymity can often enable honesty even where the statements are truthful. Many people post content anonymously because the content is potentially embarrassing to themselves rather than to others. Non-insulting anonymous content might span the range from whistleblowers who wish to reveal illegal conduct without suffering reprisal to confessional posts of those wishing to discuss sensitive subjects with a wider community without tying it to their identity, perhaps issues of physical or mental health, suicide, sexuality, frustration or anxiety with relationships with spouses, parents, children, siblings, etc. Mandating identification, even if moderated intelligently, would likely eliminate that type of positive anonymous speech as well.

[Daniel Griffin]

You might be interested in "Bowling Alone" by Robert Putnam. Not exactly the same phenomenon and of a different genre, but he brings up many similar concerns about the changing nature of group interaction.

[Dale]

The internet, just like those bathroom stalls, have a very low amount credibility. People recognize that most of the information on the internet from untrusted sources, most likely things are posted due to bouts of anger that have nothing to do with the slander itself.

Also, the idea that the ISP should be responsible for keeping track of information on all is ludicrous. The metaphor doesn't hold up. This is not like the Times being responsible for what is being posted in it's editorials, it's like the post office having to keep track of all it's shipments in case somebody sends a discriminatory letter.

Yes, I recognize that the government can try to subpoena both the post office and the newspapers, but I think that ISPs should take a page from the newspapers book and recognize if you are trying to uphold certain freedoms, you cannot divulge your sources.

Is it unfortunate that discriminatory and slanderous remarks get published on the internet (or anywhere else for that matter)? Yes, but I believe that it's unfounded to say that the cost of losing the liberties is worth the benefit of the few who would get hurt.

To quote Roosevelt, "Those who would give up essential liberty to purchase a little temporary safety deserve neither liberty nor safety."

[Bryan Hart]

Levmore responded to Dale's point by saying that, if everyone disregards the speech, what value does it have? In order for websites to maintain usefulness, some filtering becomes essential. Furthermore, although ISPs may bear some similarities to post offices, websites with comment boards are far more similar to newspapers.

Free speech arguments do not really apply because Levmore's point is that the web's users should be subject to the same substantive free speech law as users of other media. If the First Amendment protects a type of speech, that speech is protected both on and off the internet. Principle may define First Amendment doctrines, but additional protection beyond the First Amendment is a question of balancing.

Balancing brings me to Luke Gilman's point, which is essentially that Levmore underestimates the benefits. During the talk, Levmore admitted some benefits from anonymity, such as teaching evaluations or playing Second Life as a "cross-vatar" (an avatar of the opposite gender). Balancing tests are necessarily empirical, so this cannot really be answered in the abstract. The debate over the internet so far has stressed the benefits rather than the costs, so it would be a step forward to gather an accurate accounting of the costs of additional protection.

[anonymous]

Very very interesting.

This reminds me of something I heard from a colleague of Robert Solow. Apparently it is no secret that Bob Solow does not use email. His reason is something along the lines of "if the marginal cost of sending email is zero they will drop the marginal benefit to zero". Seeing what's happened with spam and how much junk there is on sites that allow users to post costlessly, I think he is onto something.

(I see the irony of me posting anonymously while the cost of deciding whether or not to filter out my post falls upon the moderator.)

On a different note, I find Levmore's analysis of gender imbalance incomplete on two counts. I just checked out a few pages on JuicyCampus and while many posts fit his description ("I was f-ing this girl on campus..."), I would venture a guess that a good number of posts attacking females are written by females, and that a very large proportion of posts are homophobic attacks on males (in fact one of the most popular tags is "gay"). Thus I would find it hard to believe that a male/female difference in preference for regulation, if any (having a larger group of vocal supporters does not mean the general male population is more supportive), is driven by victimization.

[Bryan Hart]

Anonymous: To clarify, the gender story is my own take; Levmore rests on the public choice story. The imbalance I'm pointing to is not so much in the targeters, which could be female as well as male, but in the targets. I think the popularity of the word "gay" overstates the targeting of men. First, "gay" is a default insult for men while the variety of insults for women is much greater, so the fact that "gay" is a more common word than each individual insult for women does not mean much. Second, "gay" can and is used to describe nonpeople, such as activities or institutions.

I don't think a majority of either gender has an opinion about this topic. When the internet was new, both the vocal supporters and potential detractors would have been drawn from the same group of people who are internet savvy. Since internet savvy people are disproportionately male, the group underrepresented victimization experiences, so the initial opinion formation skewed positive. This may no longer be true because the number of people who care and have an opinion about the internet has expanded, so opinion could be shifting.

[Kimball Corson]

Going back to the premise query, "Is the internet different from the bathroom stalls in some fundamental way?" I think so. The material in most bathroom stalls is usually more salacious, its audience is more narrow, its language is cruder and more outrageous, it is less reasoned and more random in content and character, and it can be scrubed clean before being read by all customary facility users. Too, decency standards do not apply, almost ever; indeed, kudos go to the most indecent. The internet is another beast altogther.

But why police the internet for decency at all? Let free speech flow as it usually does in our own lives and then let the normal consequences apply and ensue as best the law can allow. I even hate the moderation we now have here which I think dulls participation on this site considerably.

[Alvin]

Dean Levmore,

I really enjoyed your talk - didn't get much out of the question and answer period though - and I had one question and one comment.

1. How do you feel about the site "Rate My Professors"? I was surprised you didn't bring it up in your talk. Would you equate that site to end-of-class evaluations by students? It seems like a valuable site for both students and professors.

2. In your hypothetical about the victim "Amy X", you mentioned the pain caused by comments on Juicy Campus. You also mentioned how social sanctions prevent people from making nasty comments when they are not anonymous, e.g., in a classroom among peers. My question: isn't there a benefit to sites like Juicy Campus in that it might prevent people from engaging in "slutty" or otherwise inappropriate or destructive behavior? If you know your actions might be "posted" on a website, wouldn't you think twice before engaging in certain types of behavior? My point is, isn't there a benefit side that was missing from your example?

[Bryan Hart]

Alvin, I think some of your assumptions are unwarranted. I don't think that the statements can have much deterrent effect because they are mostly false; when someone posts "Amy X is a slut," it probably means that the poster doesn't like Amy X, not that Amy X actually did anything "slutty." Furthermore, the social mores on these sites are not necessarily the best; comments are often objectifying or (as noted by anonymous above) homophobic.

[Eric Goldman]

If you're going to use cost-benefit analysis, you have to include all of the socially beneficial speech that remains on the Internet solely because 47 USC 230 eliminates the pressure websites feel to take content down upon a third party's demand. Eric.

[Brian Leiter]

In reply to Professor Goldman: I take it that Dean Levmore's point is that the absence of anything like Section 230 for newspapers also comes at the cost of excluding a lot of "socially valuable speech." But we live with that cost. So why exactly should the Internet be treated differently? That's the issue, as I understand it.

[Olivia St. Clair]

I am responding to Alvin's comment. Juicy Campus' comments do not constitute helpful negative gossip for several reasons, including their frequent falseness, as Bryan suggested. Additionally, anonymity would decrease the benefit of any such gossip because anonymous sources cannot lend personal credibility to their accusations. Juicy Campus is therefore even worse than "he said, she said", since we have little information about who is saying what and why. Most importantly, the "benefit" you describe, of sanctioning women from "slutty" behavior, is deeply problematic. Why is "slutty" behavior on a woman's part undesirable and worthy of constraint? Why is it Juicy Campus commenters' privilege to judge a woman's behavior, sexual or not, as inappropriate? We know that targets of anonymous gossip on the internet of the type discussed on Juicy Campus are disproportionately female. Your comment says that a woman engaged in what you deem objectionable behavior deserves to be called out on the internet. You illustrate the problem with anonymous sites perfectly. You tend to blame Amy X rather than any parter with whom she is engaged in such behavior, based on no information, having been empowered by a social structure that allows you to act as if a female stranger's sexual behavior is within your province to regulate.

[prad]

I wonder whether Dean Levmore's new bevy of international law scholars could help him sort out the rather obvious difficulties in regulating internet speech.

In any case, I gather from Bryan's thoughtful post that the proposed solution relies on an ability subpoena internet service providers. Any thoughts on service of process on ISPs in England? In the more oxymoronically named Sealand?

I can't tell whether enforcement problems help or hurt Dean Levmore's case (in the long view), since one of the clearer answers to Professor Leiter's question would be the intra-medium alternatives available to those sophisters and reprobates rejected from newspaper editorial publication due to the lack of a 47 USC 230-type safe harbor. Subject to some restraints on the margins, the law directs them to become anonymous pamphleteers. Assuming perfect enforcement, Dean Levmore's proposal opens no such alternative method of using the internet medium, and there's nowhere else (on the web) for the "socially valuable speech" to go.

I should think that those supporters of Publius at the law school would take issue with that consequence.

[rtm]

1. If I follow the discussion correctly, the Internet (a) currently facilitates a greater degree of freedom in posting comments (ideas perhaps) through a combination of anonymity and relative lack on content restrictions; (b) does not compel anyone, other than monitors, to read such comments; and (c) allows for hosts/providers to make their own decisions, given the constraints of their ability to police those who might thwart their controls, regarding content.

2. If the Internet therefore offers something beyond that currently offered by other sources of information – television, newspapers, etc. – then it provides society with an “information option” that would presumably disappear to a large degree if the law held it to the same standard as other sources of information. It seems that there should be a rebuttable presumption that people are better off with an alternative, non-compulsive (see 1(b)), source of information.

3. If we are to rebut the presumption that alternative venues for speech are socially beneficial – and in certain cases we can – then it seems that the discussion turns initially to the benefit of paternalistic intervention, which is, I suppose, the point of this discussion. Here it seems inadequate to simply point to restrictions imposed on other media outlets as proof of the value of like restrictions applied to the Internet, and I do not take any comments to be making the argument that regulation uniformity is the ultimate goal. That said, if existing regulation of other media sources rationally prohibits “bad speech,” then there is no reason to think that any properties of Internet communication purify similar “bad speech,” and the rational basis for those regulations likely holds for the Internet. Indeed, it may well be the case that Internet users, and targets of unregulated Internet comments, are more vulnerable to the harms of “bad speech” via the Internet than they would be to similar speech through other mediums – there are several arguments for and against that assertion.

4. Supposing that we side in favor of paternalistic intervention, we next need to consider costs of enforcement. And assuming, for the sake of argument, that we become comfortable with the costs of enforcing newspaper-like regulations for the Internet for U.S. ISPs [and if Wikipedia requires 75,000 registered editors, this may require enlisting most of the population to cover all of the U.S. ISPs], how to patrol non-U.S. ISPs? Would we see “off-shore ISPs” hosting the content no longer permitted through U.S. providers? If so, do we simply follow China’s lead in blocking access?

[Alvincente]

I thought this address, although interesting, was extraordinarily weak because it entirely ignored the enormous benefits of anonymous posting on the internet. I regularly read a number of very thoughtful blogs, mostly run by professors of one sort or another. The vast majority of comments are anonymous.

The reasons not to post your name on an Internet blog are very strong. The discussions are like lunch discussions at the faculty club, only instead of being evanescent, they remain in the Internet archives forever. You may later be embarrassed by comments when you change your mind or look something up, or your stated opinions may come back to haunt you years later when you apply for a job, run for public office, or seek a judgeship.

Yet this kind of frank, wide-ranging, and sometimes passionate (as well as thoughtful) discussion is extremely valuable, and is one of the most important aspects of the Internet. There is no practical way to shut off the "Juicy Campus" type of anonymity without also shutting out the anonymity of the more serious discussion sites.

By ignoring the serious uses of anonymity and focussing only on the frivolous uses, Dean Levmore reduces his talk from an examination of his proposal to a mere brief for the prosecution. I am tempted to say that his position arises from the natural suspicion that those in positions of power have for the unregulated actions of people generally.

[TGP]

Im not sure that i see the major cost of the anonymity. remember that the comments on sites like the one used are subject to major credibility problems. They are, after all, anonymous postings on the internet. Are we really advocating that we regulate all the speech on the internet based on the fact that some people are being mean and gossipy?