The Ninth Circuit Broke The Internet. So We Asked Them To Unbreak It.

from the please-fix-this dept

It is possible that if the Ninth Circuit panel truly realized how badly it messed up Section 230 it might have thought twice about it. So we’ve asked the court to give it a second thought. As did Airbnb and Homeaway, who were most immediately affected by the Ninth Circuit’s recent decision in their challenge of the Santa Monica ordinance that, like the San Francisco ordinance, and ordinances increasingly sprouting up around the country, seeks to make them liable for their users’ expression.

The problem: that’s exactly what Section 230 is supposed to prevent – holding a platform liable for user generated content that is wrongful in some way. If Santa Monica, San Francisco, and all those other cities want to make it illegal for people to list homes to rent, that’s fine. It may or may not be good local policy, but it won’t break the Internet. What breaks the Internet is when the law doesn’t just make people legally responsible for their own expression but makes the platform they used to express it liable for it too. Section 230 is supposed to prevent that, because if platforms can be held liable for all the myriad things that can be wrong with all the enormous amounts of user expression they intermediate, then they won’t be able to be platforms anymore. It will simply be too expensive to mitigate and manage this risk, at least not in a way that doesn’t result in enormous amounts of censorship of user content that isn’t even legally wrongful at all.

So Airbnb and Homeaway filed a petition for rehearing and rehearing en banc to ask the Ninth Circuit to review their case again, and last week the Copia Institute, along with the R Street Institute, filed an amicus brief in support of their petition. In our brief we reminded the court of what we have discussed here. First, that threatening platforms with liability forces platforms to have to monitor all their user expression, which may or may not even be possible, and at the expense of any monitoring that might be more effective. For instance, in this case, all these cities are asking Airbnb and Homeaway to ensure that every listing it allows to be rented be compliant with the registration requirement, but it might be better if instead they could focus their resources on building a more usable and secure platform, helping to eliminate fraud, or working to satisfy any other priority that would benefit the public more. Threatening platforms with liability for user content inevitably co-opts platforms’ resources, diverting them away from the sort of beneficial monitoring Congress tried to incentivize them to do with Section 230 and into monitoring that is solely self-protective.

Secondly, it may likely not even be possible for platforms to do enough monitoring to protect themselves. Although the Ninth Circuit’s decision spoke to the Santa Monica ordinance, there is nothing about the decision that is limited to this specific ordinance in this specific city. A core problem with the decision is the degree to which the court minimized how difficult it will be for Airbnb and Homeaway to even just monitor their user listings to see if they comply with even just this registration requirement in even just this city. But other cities now have ordinances too, thus vastly expanding the task. There is also nothing in the decision that limits what the ordinance can demand for compliance – today it may be registration, but tomorrow it might be habitability concerns, which are even more infeasible for platforms to police, or any other arbitrary policy demand. And there is nothing limiting this tearing open of Section 230’s pre-emption provision preventing local liability from being imposed on platforms for user content to just this sort of local regulation relating to short-term rental platforms. It opens the door to absolutely everything every jurisdiction everywhere can dream up to hold against platforms. There is no way for platforms to be able to successfully monitor every regulatory demand every jurisdiction can make on user expression, so they will either give up and shut down completely or adjust their practices to comply with the most restrictive jurisdiction’s demands and ultimately end up censoring an awful lot of perfectly lawful content – or both. Section 230 was supposed to prevent platforms from finding themselves in this impossible position, and our brief reminded the Ninth Circuit of this fact.

Also, as we previously pointed out, the fundamental error of the decision is that it split out facilitating the hosting of user expression from the facilitating of a transaction related to that user transaction. If this were a legitimate distinction, it would make it impossible to ever monetize one’s platform services, because every revenue transaction would always be connected to user content that could be wrongful. It doesn’t do anything to insulate platforms from the hosting of that content if it doesn’t insulate them from being able to afford to host that content. A decision like this one directly threatens the commercial viability of the Internet, which is definitely not what Congress wanted to have happen when it passed Section 230 in order expressly to protect that economic vitality.

Filed Under: 9th circuit, cda 230, santa monica, section 230, transactions
Companies: airbnb, homeaway

9th Circuit's Bad AirBnB Decision Threatens Basic Internet Business Models

from the "but-wait,-there's-more!" dept

I'm not done excoriating the Ninth Circuit's recent decision dismissing Homeaway and Airbnb's challenge of the Santa Monica ordinance that holds them liable if their users illegally list their properties for rent. As I wrote before, that's what the ordinance in fact does, even though Section 230 is supposed to prevent local jurisdictions from enforcing laws on platforms that have this effect. Perhaps this decision may not be as obviously lethal to the Internet as the EU's passage of the Copyright Directive with Articles 11 and 13, but only because its consequences may, at the moment, be less obvious – not because they stand to be any less harmful.

Which is not to say that the court intended to herald the end of the Internet. Indeed there is a somewhat apologetic tone throughout the decision, as if the court felt it had no choice but to reach the conclusion that it did. But there is also a tone of dismissiveness that runs throughout the decision as well. The court largely minimized the platforms' arguments about how the ordinance will affect them, and by ignoring the inevitable consequences thus opened the door to them, now and in the future, far beyond the facts of this particular case.

Ultimately there are (at least) two big problems with the decision. The earlier post highlighted one of them, noting how chilling it is to speech if a law effectively forces platforms to police their users' expression in order to have any hope of avoiding being held liable for it. The problem with the court's decision in this regard is that it kept [see pages 13-14, 17, 20...] incorrectly insisting, over the platforms' protest, that the Santa Monica ordinance does not force them to monitor their users' expression when, in actuality, it most certainly does.

The second major problem with the decision is that the court kept trying to create an artificial distinction between imposing liability on platforms for facilitating user expression, which the court acknowledged would be prohibited by Section 230, and imposing liability on platforms for facilitating online transactions — which, per the court, Section 230 would apparently not prevent.

As the Platforms point out, websites like Craigslist "advertise the very same properties," but do not process transactions. Unlike the Platforms, those websites would not be subject to the Ordinance, underscoring that the Ordinance does not target websites that post listings, but rather companies that engage in unlawful booking transactions. [p. 20]

Unfortunately it's a nonsensical distinction, and one that leads to an entirely unprecedented curtailing of Section 230's critical statutory protection.

If the court's reasoning were correct, then no platform that profits from transactions between users could ever have been shielded from liability. It always would have been possible to predicate their liability on the brokering of the transaction, rather than on their intermediation of the user expression behind the transaction. In reality, though, over the past two decades plenty of transactional platforms have been able to avail themselves of Section 230's protection. For instance, EBay and Amazon, like Airbnb and Homeaway, make money from the transactions that result when user expression offering something for sale is answered by users who want to buy. Yet courts – including the Ninth Circuit – have found them just as protected by Section 230 as their non-transactional platform peers. Unlike this particular Ninth Circuit panel, these other courts recognized that the liability these platforms were having to face was inherently rooted in the user's expression, and thus something that Section 230 protected them from. For the Ninth Circuit to now decide that liability for facilitating a transaction is somehow something separate from facilitating the user expression behind it strikes at the heart of what Section 230 is supposed to do – protect platforms from liability in their users' activity – and puts all these transactional platforms' continued Section 230 protection in doubt.

It also puts in jeopardy the entire Internet economy by so severely limiting the ability of platforms to monetize their services. At minimum it calls into question any monetization model that derives revenue from any transactional user expression that successfully results in a consummated deal, since platforms can now be held to account for any alleged illegality in that user expression. (Oddly, though, a platform would seem to be just fine if its users only posted poorly-crafted listings for terrible properties at unmarketable rents, even if those listings were illegal under the ordinance, because there would be no danger of a rental transaction actually resulting. But as soon as users manage to successfully articulate their offerings such that they could result in real rentals the platform would suddenly find itself on the hook.)

Instead platforms will have to support themselves in other ways, such as by being ad-supported or charging for listings. But not only do these other revenue models raise their own concerns and considerations, but given the logic of this decision it's not a certainty that they, too, won't someday be found to put the platform beyond the reach of Section 230 as well. For as long as facilitating the exchange of money is treated as something separate from the facilitation of user expression the exchange is connected to, any method of monetary exchange connected to user content that's illegal in some way could still put the platform facilitating it beyond the reach of the statutory protection. The court cited Craigslist as an example of a platform that can retain its Section 230 immunity even when its users post illegal listings. But, notably, Craigslist does not charge users to post their listings. Which leaves us with a decision where the only platforms that can be sure to benefit from Section 230 are the ones that provide their services for free, which isn't consistent with what Congress intended or the commercial potential upon which, until now, the Internet economy has depended.

Furthermore, it stands to put any platform exclusively devoted to facilitating aspects of these transactions completely beyond the protective reach of Section 230, even though any liability connected to that transaction would still be due to others' expression. In this case the court was fairly indifferent as to what a platform like Homeaway or Airbnb would need to do to cope with liability under the ordinance.

[T]he Platforms argue that the Ordinance "in operation and effect . . . forces [them] to remove third-party content." Although it is clear that the Ordinance does not expressly mandate that they do so, the Platforms claim that "common sense explains" that they cannot "leave in place a website chock-full of un-bookable listings." For purposes of our review, we accept at face value the Platforms’ assertion that they will choose to remove noncompliant third-party listings on their website as a consequence of the Ordinance. Nonetheless, their choice to remove listings is insufficient to implicate the CDA. [p. 14-15]

Or that the platforms' attempt to avoid liability might result in undue censorship.

Moreover, the incidental impacts on speech cited by the Platforms raise minimal concerns. The Platforms argue that the Ordinance chills commercial speech, namely, advertisements for third-party rentals. But even accepting that the Platforms will need to engage in efforts to validate transactions before completing them, incidental burdens like these are not always sufficient to trigger First Amendment scrutiny. [p. 20-21]

These are significant concerns, however. It is a functionally impossible task the court sets for them, to put platforms in the position of needing to review and remove just the right amount user expression in order to protect themselves. But this sort of case-by-case, listing-by-listing censorship is at least theoretically within the power of a platform facilitating the expression itself. Whereas it's not at all within the power of platforms like payment providers that never touch the original user speech. Instead they will be left with a stark choice: leave open the firehose to process all transactions, and thus risk being liable for any transaction related to any illegal listing that comes through, or turn off their service entirely to any service that cannot guarantee to them that every transaction arising from every user listing will be legally compliant with every possible law. No platform can make that promise, of course, and it will inevitably result in the removal of substantial lawful content to even try.

True, the Santa Monica ordinance itself does not invite this full parade of horrors. It is but one ordinance, for one city, with a regulatory ask of the platform that the court somehow seems to think is easy to meet. But this decision and its contorted rationale opens the door to plenty more ordinances, from plenty more jurisdictions, with plenty more regulatory demands. In and of itself it is plenty onerous, and it invites even worse.

Filed Under: 9th circuit, business models, cda 230, santa monica, section 230, transactions
Companies: airbnb, homeaway

Ninth Circuit Tells Online Services: Section 230 Isn't For You

from the practical-effect dept

Last year we wrote about Homeaway and Airbnb's challenge to an ordinance in Santa Monica that would force them to monitor their Santa Monica listings to ensure they were legally compliant. The Santa Monica ordinance, like an increasing number of ordinances around the country, requires landlords wanting to list their properties on these services to register with the city and meet various other requirements. That part of the ordinance is not what causes concern, however. It may or may not be good local policy, but it in no way undermines Section 230's crucial statutory protection for platforms for Santa Monica officials to attempt to hold their landlord users liable if they go online to say they have a non-compliant rental listing.

The problem with the ordinance is that it does not just impose liability on landlords. It also imposes liability on the platforms hosting their listings. The only way for them to avoid that liability is to engage in the onerous, if not outright impossible, task of scrutinizing whether or not the listings on their platforms are legal. Which is exactly what Section 230 exists to prevent: forcing platforms to monitor their users' speech for legality, because if they had to police them, they would end up facilitating a lot less legitimate speech.

Yet that's what the Ninth Circuit decided to let Santa Monica do – force platforms to monitor their user-generated speech – in a decision earlier this week upholding the district court's refusal to enjoin the ordinance.

Of course, that's not how the court saw it, however. To the court, platforms weren't being forced to police the speech they hosted. They were merely obligated to police the rental transactions they facilitated.

[T]he Ordinance does not require the Platforms to monitor third-party content and thus falls outside of the CDA’s immunity … [T]he only monitoring that appears necessary in order to comply with the Ordinance relates to incoming requests to complete a booking transaction—content that, while resulting from the third party listings, is distinct, internal, and nonpublic. [p. 13-14]

However this is a distinction without a difference.

As we pointed out in the amicus brief the Copia Institute filed in support of Homeaway and Airbnb, these listings are indeed user-generated speech. It may be speech that's extremely limited in scope, little more than "I have housing to rent," but it is still user speech that, per the ordinance, may not always be legal to say. The problem is that this ordinance in effect is all about passing on liability to the platform if they allow this speech to be illegally said, which is no different than trying to pass on liability to a platform for any other speech its users may illegally say.

Yet in its decision the court insisted that platform liability attaches to something entirely apart from its role as a platform facilitating user speech:

Similarly, here, the Ordinance is plainly a housing and rental regulation. The “inevitable effect of the [Ordinance] on its face” is to regulate nonexpressive conduct—namely, booking transactions—not speech. [p. 19-20]

It went on to declare that the ordinance in no way forces platforms to monitor user content:

Contrary to the Platforms’ claim, the Ordinance does not “require” that they monitor or screen [listings]. It instead leaves them to decide how best to comply with the prohibition on booking unlawful transactions. [p. 20]

At every step in its reasoning it kept treating the ordinance as something wholly apart from an ordinance impacting speech:

Nor can the Platforms rely on the Ordinance’s “stated purpose” to argue that it intends to regulate speech. The Ordinance itself makes clear that the City’s “central and significant goal . . . is preservation of its housing stock and preserving the quality and nature of residential neighborhoods.” As such, with respect to the Platforms, the only inevitable effect, and the stated purpose, of the Ordinance is to prohibit them from completing booking transactions for unlawful rentals. [p. 20]

But no amount of handwaving by the court to try to focus on the financial transaction between landlord and renter, or insistence that this ordinance doesn't force platforms to monitor user-generated speech, will change the basic reality that it does indeed force platforms to do exactly that: police user speech for legality in order to avoid liability arising from that speech. It is exactly the sort of situation Section 230 was intended to forestall because of the inevitable chilling effect fear-driven platform monitoring obligations have on online speech and innovation.

The court seemed to try to justify its contorted reasoning by noting that because "brick and mortar" businesses have to comply with all sorts of local regulations, Internet businesses also should have to.

We have consistently eschewed an expansive reading of the statute that would render unlawful conduct “magically . . . lawful when [conducted] online,” and therefore “giv[ing] online businesses an unfair advantage over their real-world counterparts.” For the same reasons, while we acknowledge the Platforms’ concerns about the difficulties of complying with numerous state and local regulations, the CDA does not provide internet companies with a one-size-fits-all body of law. Like their brick-and-mortar counterparts, internet companies must also comply with any number of local regulations concerning, for example, employment, tax, or zoning. [p. 16]

But this thinking fails to recognize the unique differences between brick and mortar businesses and Internet business, differences that help explain why it is so important to give Internet businesses this vital protection. After all, a brick and mortar store only has to comply with the laws of the jurisdiction where the store is located – as Internet platforms also need to, in the finite number of places where they have a physical or corporate presence. But when it comes to their online presence, an Internet business is everywhere and thus theoretically exposed to the laws of every single jurisdiction, no matter how onerous these laws are, or how much they may conflict with any other's.

Because while perhaps the Santa Monica ordinance may not be too onerous for the platforms to comply with in and of itself, Santa Monica is but one city, yet the Ninth Circuit has now given the green light to every other city in every other state to come up with their own ordinances that will similarly force platforms to monitor user content. As Congress feared in 1996 when it passed Section 230, this decision now invites platforms to divert resources better spent elsewhere, overly censor user speech, withdraw from entire markets – even those that might prefer to have these services available – or risk being bankrupted by an infinite number of local jurisdictions pulling them in every possible direction.

This result is chilling not just to these platforms but to any other innovative service, especially if the service has any effect in the offline world, as so many do, or facilitates economic transactions between users, as so many also do. If bearing these indicia are enough to cause a platform to lose its Section 230 protection, then few will be able to retain it.

Filed Under: 9th circuit, california, cda 230, housing, intermediary liability, internet, platforms, santa monica, section 230, speech
Companies: airbnb, homeaway

Airbnb, Homeaway, And The Importance Of Holding The Line On Section 230

from the back-to-basics dept

SESTA has done enormous damage to the critical protection Section 230 affords platforms – and by extension all the Internet speech and online services they facilitate. But it's not the only threat: courts can also often mess things up for platforms by failing to recognize situations where Section 230 should apply and instead allowing platforms to be held liable for how their users have used their services.

Which leads to the situation Airbnb, Homeaway, and other such platforms find themselves in. Jurisdictions unhappy with some of the effects short-term rentals have had on their communities have taken to passing regulations designed to curb the practice. Whether or not it is good policy to do so is beyond the scope of this post. If some local jurisdictions want to impose liability on their residents for renting out their homes – and not all of them do – it's between them and their voters.

The problem arises when the regulations they come up with don't just target people renting their homes, but also target the online platforms that facilitate these transactions. These ordinances effectively create liability for platforms arising from content generated by others, which is a regulatory practice that Section 230 prohibits.

So Airbnb and Homeaway have started pushing back on these ordinances, first in San Francisco and now in Santa Monica. Unfortunately both efforts to enjoin them have resulted in federal district court decisions saying that Section 230 does not shield them from their reach, meaning that these local jurisdictions are fully able to hold these platforms liable if people use them to rent homes they aren't supposed to. The decision about the Santa Monica ordinance is now before the Ninth Circuit, and last week I wrote a brief for the Copia Institute explaining why it should find that Section 230 indeed prevents these ordinances from imposing liability on these platforms. It was important to say so, not just to support Airbnb and Homeaway, but because if Section 230 can't apply to them, then it won't be able to apply to a lot of other platforms that depend on it.

The crux of the problem appears to stem from courts not seeing how what is at stake in these cases is actually speech, perhaps because the kind of speech sites like Airbnb and Homeaway intermediate is so specific. But even if the only expression these platforms intermediates is, "I have a home to rent," it's still speech, speech created by someone other than the platform, and Section 230 therefore still applies. There is no language in Section 230 that would require a platform to intermediate lots of different kinds of expression in order to be entitled to the statute's protection. Many platforms are extremely specialized in the type of expression they intermediate, often because that's what makes them useful and effective as services, and all are equally entitled to the statute's protection.

The fact that the specific speech being intermediated is transactional in nature seems to be what's confusing the courts, especially given that these sites often make money by taking a cut of the transactions that are successful. The court addressing the Santa Monica ordinance recognized that a site like Craigslist, which also hosts "I have a home to rent" speech (among other types of speech), would not be affected by the ordinance because it doesn't make money when "I have a home to rent" speech results in a rental. But there is no reason that these platforms should be treated any differently. Section 230 applies regardless of how a platform makes its money. There's no requirement in the statutory language that a platform profit only in certain ways – in fact, if anything the statute encourages platforms to be innovative so that the public can continue to benefit from their services. And for good reason: think about platforms like eBay, which also profit when "I have a thing to sell" speech finds an audience who wants to buy it. If Section 230 protection could be withheld from all platforms that make money from consummated transactions it would be more than just the Airbnb and Homeaway who would be in trouble.

The only relevant question to ask in considering whether Section 230 applies is who created the content that is potentially wrongful. In the case of Airbnb and Homeaway it is their users. After all, there's nothing inherently wrongful about saying, "I have a home to rent." Whether it is wrongful depends entirely on whether the user is allowed to rent it per local law. Liability should therefore remain entirely with the user who is the one who imbued it with its wrongness. Particularly because it is often not practical, or even possible, for platforms to police all the content passing through them. Even if they had the resources to examine the volume of user-generated content that passes through their systems they may not have the ability to know which, if any of it, was wrongful. Thus if platforms could be forced by any particular jurisdiction to try to police it anyway, in order to stave off potentially expensive liability, it would invariably chill their ability to provide their services – including in other jurisdictions.

Which is also why Section 230 includes a pre-emption provision, so that no particular jurisdiction can get to decide for any other one what Internet speech and services people can benefit from in these other places. Without that provision the jurisdiction with the most restrictive laws would otherwise get to impose its policy choices on any other jurisdiction the service now shaped by these policies could reach, which, in the case of an Internet service, is every single one of the thousands and thousands of state and local jurisdictions nationwide.

Filed Under: cda 230, home rentals, intermediary liability, santa monica, section 230, speech
Companies: airbnb, homeaway