Roommates.com Denied 230 Immunity by Ninth Circuit En Banc (With My Comments)
By Eric Goldman
Fair Housing Council of San Fernando Valley v. Roommates.com, LLC, 2008 WL 879293 (9th Cir. April 3, 2008)
The Ninth Circuit has issued its en banc ruling in the Fair Housing Councils v. Roommates.com case upholding the prior ruling of the three judge panel. The opinion concludes that Roommates.com does not qualify for 47 USC 230 immunization for some of its activities and sends the case back to the trial court for further litigation on other substantive questions, such as whether Roommates.com’s behavior actually violated the Fair Housing Act.
The en banc panel voted 8-3 to reverse the district court. Judge Kozinski wrote the majority opinion, which was not surprising given that Kozinski had also written the lead opinion in the three-judge panel ruling. Judge McKoewn wrote the concurring/dissenting opinion.
Before I get into the opinion’s substance, I want to note that the new opinion shows that the en banc process worked well here. Although I don’t think that Kozinski’s opinion is as clear as he thinks it is (for example, in FN 39, he says “there can be little doubt that website operators today know more about how to conform their conduct to the law than they did yesterday”), this opinion is more careful and has less ambiguity than his prior opinion. I still don’t agree with Kozinski’s conclusions or arguments, but this opinion was written with a lot more precision and, as a result, probably will have fewer adverse collateral consequences than his prior opinion. Kudos to Judge Kozinski for responding to the feedback he got and to the Ninth Circuit for using the en banc process to improve the clarity of its legal rules.
Substantively, Kozinski says that Roommates.com does not qualify for 230 immunization for (1) the questions it poses to users (such as asking racial preferences), (2) users’ profile answers made using pull-down menus (at least to the extent that the site required answers before the user could proceed), and (3) its search system and personalized email system that display profile results (and, by implication, suppressed other results) based on impermissible criteria. However, 230 immunized Roommates.com’s unstructured “Additional Comments” data field, even if impermissible sentiments were expressed in it.
There is lots of interesting language and examples in both the majority and the dissent, and it’s worth the read. I’m focusing this post on a few themes in the majority opinion that I found particularly noteworthy:
47 USC 230 was enacted in 1996 during the height of “cyberspace exceptionalism,” the belief that the Internet was unique/special/different and therefore should be regulated differently. 47 USC 230 is a flagship example of such exceptionalism. It creates rules that really differ between the online and offline worlds, such that publishing content online may not create liability where publishing the identical content offline would. The medium matters.
Despite Congress’ obvious intent, the majority opinion bristles with antipathy towards cyberspace exceptionalism. In numerous places, it expresses skepticism that offline rules should not apply verbatim to online behavior–even though that’s exactly what 230 does in some cases. Kozinski’s not the first judge to rankle at the policy implications of Congress’ cyberspace exceptionalism, but he may have let this hostility override other analytical considerations.
As result, I think the majority opinion reflects an implicit normative judgment that the Fair Housing Act should trump 47 USC 230, which distorts the statutory analysis. (To be clear, the dissent makes this point too, so I’m not the only one who feels this way). I understand why the majority would go this direction; anti-discrimination goals are normatively very important. At the same time, 230 codifies some very important norms too.
Like his last opinion, Kozinski’s opinion repeatedly indicates that unstructured searches pose fewer problems for 230 than structured ones. Specifically, the majority opinion repeatedly says that “neutral tools” (a term used 5 times but never defined) fully qualify for 230 protection, even if these search tools are misused for impermissible searches. Thus, the opinion explicitly says that general purpose search engines like Google can qualify for 230 even if someone uses it to engage in searches that violate the Fair Housing Act.
The majority opinion further tries to narrow the offending behavior significantly. It says that pull-down menus aren’t inherently problematic. For example, “A dating website that requires users to enter their sex, race, religion and marital status through drop-down menus, and that provides means for users to search along the same lines, retains its CDA immunity insofar as it does not contribute to any alleged illegality.” [So, if someone entered an age below 18 and the lawsuit was over a subsequent sexual assault or statutory rape, no immunity?] It also says that a site that enables searches via “user-defined criteria” would retain the immunity [not sure what that means]. Even so, the ambiguities in this discussion may unfortunately dissuade some websites from providing structured searches of user-supplied content.
The Role of Passivity
Prior to the passage of 230 in 1996, the conventional Cyberlaw wisdom was that an online entity wasn’t liable for user content only if it acted like a “passive conduit.” If the entity actively managed its user-supplied content, it could not avoid liability for that content.
We had thought 230 changed that paradigm and eliminated any differences in legal treatment between passive conduits and active content managers. Unfortunately, this opinion raises questions about that. The majority opinion uses the term “passive” 7 times in its opinion, each time implying that passivity contributes to the immunization.
At the same time, the majority opinion also contains some contrary language suggesting that active involvement still earns 230 protection. The majority specifically gives an example where a website that edits user content doesn’t lose the immunity unless the editing changes the content substantively:
A website operator who edits user-created content—such as by correcting spelling, removing obscenity or trimming for length—retains his immunity for any illegality in the user-created content, provided that the edits are unrelated to the illegality.
This sounds like a reasonably active involvement with user content. So what does passivity mean, and why did the opinion consistently refer to it?
What Does this Opinion Mean?
Overall, the majority tried hard to communicate that the rule it articulated here should apply narrowly to a limited number of sites: “The message to website operators is clear: If you don’t encourage illegal content, or design your website to require users to input illegal content, you will be immune.” The opinion also tries to signal that this ruling isn’t a major shift away from 230:
Websites are complicated enterprises, and there will always be close cases where a clever lawyer could argue that something the website operator did encouraged the illegality. Such close cases, we believe, must be resolved in favor of immunity, lest we cut the heart out of section 230 by forcing websites to face death by ten thousand duck-bites, fighting off claims that they promoted or encouraged—or at least tacitly assented to—the illegality of third parties.
These broad pronouncements are helpful, and I hope future judges take them to heart. Unfortunately, it’s virtually impossible to articulate in crystal-clear terms why Roommates.com crossed the line while many other websites with similar user interactions still qualify for 230. As a result, despite the broad pro-230 philosophical statements, I’m fairly confident that lots of duck-biting plaintiffs will try to capitalize on this opinion and they will find some judges who ignore the philosophical statements and instead turn a decision on the opinion’s myriad of ambiguities. Ultimately most courts will reach the right results, but only after significant transaction costs, and even so I think erroneous judgments are inevitable. At least this will create a steady stream of appeals that let Kozinski further amplify his views on 230.
While I expect plaintiffs to be emboldened by the lawsuit and to look at the Ninth Circuit as the new venue of choice (replacing the Seventh Circuit after the Craigslist case didn’t turn out to be plaintiff-favorable), I also wonder if this case will spur plaintiffs as much as the recent Mazur case. That case highlights the risk that websites can be attacked for their marketing statements even if the representations are rendered untrue by third party actions or conduct. So, for example, in theory plaintiffs could attack a site like Roommates.com for any marketing statements that say or imply that using its tools are legal. If these kind of “back-door” attacks on a site aren’t covered by 230 (and there may be good reasons why they aren’t), then plaintiffs won’t have to try to exploit ambiguities in the Roommates.com opinion to get their desired results. If so, while this opinion may not be a great one for defendants, there may be even darker clouds on the horizon.
* Dan Solove (this used to link to Concurring Opinions)
UPDATE: Michael Risch writes me:
You said: “Unfortunately, it’s virtually impossible to articulate in crystal-clear terms why Roommates.com crossed the line while many other websites with similar user interactions still qualify for 230.”
Here’s my attempt at clarity: Roommates.com REQUIRES people to state their race, and thus is a co-developer of what is published on its site. The other sites likely did not have mandatory choices with mandatory publication.
Thought experiment – the comment section on my blog allows for free responses, but at the bottom is a drop-down box, in which every user must make a choice in order to post the comment: “I hate jews”, “I hate women”. When the comment appears, the user’s text, and the text of the selection shows up in the comment. Sure, the user selected the statement, but I’m the one who came up with the idea and forced the user to select on or the other, and then forced the text to display. I have to think that is a development of the content.
Of course, the extension of this is that epinions now gets sued where it requires a 1 star to 5 star rating. That doesn’t seem right, but maybe that’s on the merits and not on the immunity. You would, I imagine, say it SHOULD be about the immunity, but I’m not so sure.
I think the even less clear question is what if it is NOT mandatory, and merely suggestive.
As for the searching aspect, here is the attempt at clarity: if Roomates.com had a button that said “find white roommates” it would be the speaker of a discriminatory search practice. Allowing a choice of white, black, etc. as a search item is simply an expansion of that.
The case library:
* Amicus brief from various news organizations
* The Ninth Circuit order granting the en banc hearing
So what does a duck bite look like? See here. Quack!