Properly interpreted, Section 230 is not “anticompetitive” because it does not advantage one business or set of businesses over others that operate in the same market. Its core purpose is to allow online platforms to host and moderate speech under conditions that differ from those in the offline world. There have been attempts to use Section 230 to advantage some online businesses over direct competitors, for example, in the home rental and retail markets. But courts have generally concluded that Section 230 does not preclude regulation of business activities that do not themselves involve the publication or moderation of third-party speech, even though they may relate to it in some way.
By John Bergmayer1
Section 230 of the Communications Act2 is a simple but wide-reaching law with significant competitive implications. Its core function of facilitating the hosting and moderation of third-party speech allows large and small online platforms and services to exist that otherwise would face a more challenging, or impossible, legal environment.
Among civil society groups, some argue that Section 230 allows platforms to serve as venues for free speech, and protects voices that would otherwise be silenced; some see it as a tool that allows hate, harassment, and harm to flourish online, giving victims no practical recourse.
Businesses whose operating models are challenged by the internet or by major platforms see Section 230 as a form of subsidy, allowing online services to avoid the typical costs of doing business in terms of tort liability, and allowing them to escape the harms their practices and business models create. Internet platforms, of course, see Section 230 that allows pro-competitive entry into the market.
These issues relating to free expression and the health of the information ecosystem take up a great deal of my work as a consumer advocate. But in this article I want to address the narrower question of whether Section 230 is anticompetitive, and what that would mean.
The argument set out in this piece is that, when applied to its core function of facilitating online speech, Section 230 is competitively neutral, or even pro-competitive. This does not mean it’s a perfect law. But its competitive effects should be properly understood.
However, if allowed to shield behavior beyond its core function of facilitation the publication and moderation of speech, it can become anticompetitive, by granting advantages to business that operate according to some online business models that are not available to their offline counterparts, and that have no, or an attenuated connection to facilitating speech. When interpreting Section 230, judges should be sure not to extend the law beyond its intended effects, and when considering Section 230 reforms, policy makers should think about its competitive effects in this light.
II. SECTION 230’S IMMUNITIES FOR HOSTING AND MODERATING SPEECH ARE NOT ANTICOMPETITIVE
Section 230’s core function is to allow online services to host third-party speech without facing liability as a “publisher or speaker,” while strengthening their discretion (which the First Amendment also gives them) to keep up, take down, restrict, or promote speech on their platforms according to whatever criteria they see fit. This is in some ways the most controversial part of Section 230, and it has had significant effects on the media landscape, democracy, and public discourse — all of which are beyond the scope of this short article.
Section 230 applies to all online services equally, including the online components of business that also have offline operations, and does not create an incentive for one approach to hosting or moderating content over another. If the law had continued to develop in the direction some feared at the time of the passage of Section 230, sites might have faced a difficult trade-off: either adopt a hands-off approach to content moderation where you are liable, at most, for removing specific pieces of potentially unlawful content that you are alerted to, or heavily moderate your site, but take full responsibility for the content that you leave up.3 Section 230 takes the law out of the equation — sites can heavily moderate content according to whatever criteria they want, or not, without liability questions pushing them toward one approach or another — just economic, social, and other considerations.
Section 230 allows for the emergence of business models, such as social media, that “compete for attention” with traditional media. To the extent that Section 230 allows for business to exist that otherwise would not it certainly has an effect on competition. It might even have a bad effect on competition to the extent that certain online services might be socially harmful.4 But this is not the same thing about a law having an unfair effect on competition, which I would view as giving an advantage to one company, and not another, even though they effectively do the same thing.
To state the obvious, online and offline media operate in a different way (and, in antitrust parlance, compete in a different product market). Traditional print and broadcast media have limited space, as measured in column inches, or the number of hours in a day. Most of the content they publish is their own, first-party content, which they are liable for (just as online service are). The burden of screening the limited amount of third-party content they might print or air — callers to a talk show, invited guests, or letters to the editor — do not appear to be particularly onerous, nor does liability for this content appear to be a particular onerous burden for traditional media.
Online services typically work in a different way. They publish content with no or limited pre-screening, and moderate after the fact. This allows for discussions between users, and for people to use platforms to publish their views that would never find space on traditional media. For some platforms, too, the scale at which they operate has a quality of its own. A broadcast network airs 24 hours of content per day, much of which it created itself. It is not unreasonable for it to be legally liable for all 24 hours of this content. By contrast, about 30,000 hours of video are uploaded to YouTube every hour, or 720,000 hours per day. A legal regime that held YouTube liable according to traditional “publisher” standards for every minute of that video might simply make YouTube impossible. And this would hold true even if YouTube were broken up into dozens of smaller companies.
Again, one might reasonably believe that business models like YouTube’s are problematic and that Section 230 should be reformed so that it no longer enables them. But there is no reason to view such a critique through a lens of unfair competition.
III. SECTION 230 DOES NOT PREVENT STATES AND LOCALITIES FROM REGULATING BUSINESSES THAT OPERATE IN THEIR JURISDICTIONS
Section 230 is a short law, with sweeping words. Many plaintiffs’ attorneys and other advocates argue that judges have made the law stronger than it was intended. At the same time, judges are naturally skeptical of creative pleadings that try to find a back door to holding platforms liable for activities that Congress clearly immunized.
However, litigation around local laws seeking to regulate online platforms that allow users to rent their properties for short times such as AirBnB and HomeAway showed that Section 230 does not stand in the way of efforts to regulate online marketplaces in ways that are similar to how their offline counterparts are regulated.
Attempts to regulate marketplace listings themselves would be preempted by Section 230. If a community passes a law requiring certain listings to be taken down, and a platform does not comply, holding it liable for this would be holding it liable “as a publisher.”
Many local communities wish to regulate online rental marketplaces like AirBnB in the same way they regulate hotel and home rentals. Naturally, they are less interested in regulating advertisements for rentals than rentals themselves. At the same time, it is more efficient for them to attach new legal obligations on rental platforms themselves, rather than trying to enforce against individual renters. The case law confirms that they have the right to do this, if they prohibit the transaction associated with the rental itself (business conduct), as opposed to the listing (speech).
Rental platforms argued that ordinances attaching liability to completing rental transactions were preempted by Section 230 since as a practical matter they affected what listings they would take down and what they would keep up. But courts have rejected this line of reasoning. As the United States District Court for the Northern District of California wrote,
But the Ordinance does not threaten the liability plaintiffs fear. As the text and plain meaning of the Ordinance demonstrate, it in no way treats plaintiffs as the publishers or speakers of the rental listings provided by hosts. It does not regulate what can or cannot be said or posted in the listings. It creates no obligation on plaintiffs’ part to monitor, edit, withdraw or block the content supplied by hosts. To the contrary, as San Francisco has emphasized in its briefs and at oral argument, plaintiffs are perfectly free to publish any listing they get from a host and to collect fees for doing so — whether the unit is lawfully registered or not — without threat of prosecution or penalty under the Ordinance. The Ordinance holds plaintiffs liable only for their own conduct, namely for providing, and collecting a fee for, Booking Services in connection with an unregistered unit.5
A similar ordinance was challenged and made its way to the Ninth Circuit, which similarly observed:
Here, the Ordinance does not require the Platforms to monitor third-party content and thus falls outside of the CDA’s immunity. The Ordinance prohibits processing transactions for unregistered properties. It does not require the Platforms to review the content provided by the hosts of listings on their websites. Rather, the 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… [I]t is not enough that the third-party listings are a “but-for” cause of such internal monitoring. The text of the CDA is “clear that neither this subsection nor any other declares a general immunity from liability deriving from third-party content.” Barnes, 570 F.3d at 1100. To provide broad immunity “every time a website uses data initially obtained from third parties would eviscerate [the CDA].” Barnes, 570 F.3d at 1100. That is not the result that Congress intended.6
These cases demonstrate a clear, and administrable principle. While Section 230 shields platforms from liability for hosting third-party content, it does not shield their business models. The business activities that platforms engage in as a result of user-submitted material are not shielded by 230. 230 shields their ability to host the material, but not to monetize it in any way they see fit.
This principle is also applicable to a more complicated situation — whether, and how, online marketplaces should be considered the “sellers” of third-party goods, a legal designation that could make the marketplace liable for any defects in or damages caused by the products they sell.
IV. SECTION 230 DOES NOT PRECLUDE HOLDING ONLINE MARKETPLACES LIABLE AS THE “SELLER” OF GOODS
The question of whether to treat Amazon (or other online marketplaces) as “sellers” of third-party goods has been actively litigated in the past few years. The question does not always turn on whether such a claim would be preempted by Section 230. For example, in Eberhart v. Amazon, 325 F. Supp. 3d 393, 400 n.5 (S.D.N.Y. 2018), the District Court disposed of it in a footnote. However that same court cited a series of opinions in other jurisdictions that form an “emerging consensus against construing Amazon as a ‘seller.’” 400.
Whether or not such a consensus exists or is wise is a separate question than whether it’s a result of Section 230. While synthesizing all the various cases applying the state law of seller liability and how it intersects with Section 230 for a through line may be beyond the scope of this article, looking at a few of the cases shows that, while Section 230 might indeed seem like a formal block to liability at times, this may only be in circumstances where the underlying course of conduct by the platform would not have been enough to trigger liability as a seller in the first place.
First, a somewhat easy case. In Erie, the Fourth Circuit determined, using logic analogous to the rental marketplace cases, that liability as a “seller” of a good is not precluded by Section 230, since being a seller involves more than merely hosting third-party speech. The court explained,
The products liability claims asserted by Erie in this case are not based on the publication of another’s speech. The underpinning of Erie’s claims is its contention that Amazon was the seller of the headlamp and therefore was liable as the seller of a defective product. There is no claim made based on the content of speech published by Amazon — such as a claim that Amazon had liability as the publisher of a misrepresentation of the product or of defamatory content….While the Communications Decency Act protects interactive computer service providers from liability as a publisher of speech, it does not protect them from liability as the seller of a defective product.7
Oberdorf illustrates this as well. Interestingly, however, this same principle animates both the district court opinion, and the appellate court opinion, which reversed it.
The court first characterized Amazon’s conduct, concluding that it does not qualify as a “seller” under Pennsylvania law. Analogizing to a previous case involving an auctioneer who was found not to be a “seller,” the court reasoned as follows:
Like an auctioneer, Amazon is merely a third-party vendor’s “means of marketing,” since third-party vendors — not Amazon — “choose the products and expose them for sale by means of” the Marketplace. Because of the enormous number of third-party vendors (and, presumably, the correspondingly enormous number of goods sold by those vendors) Amazon is similarly “not equipped to pass upon the quality of the myriad of products” available on its Marketplace. And because Amazon has “no role in the selection of the goods to be sold,” it also cannot have any “direct impact upon the manufacture of the products” sold by the third-party vendors.
The Amazon Marketplace serves as a sort of newspaper classified ad section, connecting potential consumers with eager sellers in an efficient, modern, streamlined manner.8
After determining that Amazon’s conduct did not qualify it as a “seller” under state law, the court dismissed of additional claims as precluded by Section 230.
Although the Complaint frames those claims broadly, it is clear from the Oberdorfs’ papers that they are, in fact, attempting to hold Amazon liable for its role in publishing an advertisement for The Furry Group’s product. In other words, the Oberdorfs are attempting to “treat[ Amazon] as the publisher or speaker of … information provided by” The Furry Group. Therefore, these claims are barred by § 230 of the CDA.9
But the main claim — that Amazon should be considered a seller — was not precluded by Section 230. It was precluded by the court’s conclusion that Amazon is not a seller to begin with.
By contrast, the Third Circuit found that Amazon, in this circumstance, was a seller under Pennsylvania law, “based on its role in effectuating sales of physical products offered by third-party vendors,” 143 n.11, agreeing with the plaintiffs that seller liability would be based on Amazon’s “direct role in the actual sale and distribution of the defective product,” 152, and not its editorial role as a platform.
While we recognize that Amazon exercises online editorial functions, we do not agree that all of Oberdorf’s claims seek to treat Amazon as the publisher or speaker of information provided by another information content provider. As previously discussed, Amazon is a “seller” of products on its website, even though the products are sourced and shipped by third-party vendors such as The Furry Gang. Amazon’s involvement in transactions extends beyond a mere editorial function; it plays a large role in the actual sales process. This includes receiving customer shipping information, processing customer payments, relaying funds and information to third-party vendors, and collecting the fees it charges for providing these services.
§1 Oberdorf v. Amazon.com, Inc., 930 F. 3d 136, 152-53 (3d. Cir. 2019). Therefore, to the extent that Oberdorf’s negligence and strict liability claims rely on Amazon’s role as an actor in the sales process, they are not barred by the CDA.10
While the court did find that a “failure to warn” claim was precluded by Section 230, the primary product liability claim could proceed.
As far as Section 230 is concerned, it is possible to reconcile the district court and Third Circuit opinion. In the Oberdorf District Court opinion, Amazon’s connection to selling the item was seen as so minimal that under Pennsylvania state law, not only would it not be considered a “seller,” but the the court was able to characterize the product listing as an “advertisement.”
Under this view, Amazon’s connection with the sale was primarily as a publisher of information, and the fact that it offered payment processing services was not enough for it to be a “seller” under Pennsylvania law. Given this perceived attenuated relationship between the product and the buyer, holding Amazon liable in any way would amount to holding it liable for publishing speech, and is therefore preempted by Section 230.
The Third Circuit’s disagreement with the district court was as much a factual issue as a legal one. Emphasizing that Amazon “plays a large role in the actual sales process,” meant among other things that Amazon was a seller, and if it was a seller, then holding it liable for that was not the same as holding it liable merely for publishing an “advertisement.”
In other words, if Amazon is considered a seller under state law, almost by definition, its involvement in the sales process is more involved than a hosting mere product listing or something like a classified ad. Thus holding it liable would not be precluded by Section 230.
By contrast, if a platform’s relationship to a transaction is so attenuated that it would not be considered a seller under state law, then Section might also block the claim. But the claim would likely fail anyway, as there is no underlying liability to begin with.
It might also be the case that a platform’s connection with a transaction is involved enough to overcome Section 230, but still not enough to be considered a seller under relevant state law. This was actually the case with Erie. The product in dispute was part of Amazon’s fulfillment program. Thus, in addition to being listed on Amazon’s website, the products were stocked and shipped by Amazon. The court considered this involvement with the transaction to be enough that any liability would not necessarily hinge on liability for third-party content. Yet this still wasn’t enough for Amazon to be considered a “seller” under Maryland law. But it was enough for Section 230, at least, to not preempt the claim.
While Section 230 complicates the analysis, at least as far as these three opinions are concerned, it doesn’t change the outcome. Section 230 blocks claims that hold a platform liable merely for hosting content, but if it’s merely hosting content, state law is unlikely to hold it liable as a seller to begin with. If a platform does enough to be considered a seller under state law, then holding it liable for those same actions is not blocked by 230, since it is being held liable for a course of conduct that is distinct from its role as a publisher, and which leaves its editorial prerogatives intact.
A broad law like Section 230 does carry the risk of unfairly distorting the marketplace. But Section 230’s core function of facilitating the hosting and moderation of third-party speech by online platforms does not give one set of actors an unfair advantage over their direct competitions, though of course it can still be criticized for other reasons. While platform defendants have invoked Section 230 in context where, if successful, the law would give them an unfair advantage over direct competitors, courts have proven able so far to distinguish holding a company liable for hosting and moderating speech — which 230 continues to provide immunity for — and from its other business conduct, which is outside the scope of Section 230 and subject to the usual police powers of state and local authorities.
1 John Bergmayer is Legal Director at Public Knowledge.
3 In many ways, for example the right of platforms to take down content they find objectionable, Section 230 merely adds procedural protections to existing First Amendment protections. It is impossible to know how the common law would have developed absent 230, though it is unlikely that Stratton Oakmont v. Prodigy Services, 23 Media L. Rep. 1794 (N.Y. Sup. Ct. 1995), the case frequently cited as a motivation for the passage of 230, a poorly-reasoned and flawed decision from a lower court in New York holding that if platforms moderate content they become responsible for everything they keep up, would have become the law of the land.
4 Arguments about intellectual property rights, real or proposed, for example that platforms unfairly compete with traditional media by linking to and excerpting them, are interesting, but not part of a Section 230 discussion, as Section 230 does not apply to IP claims.
5 Airbnb, Inc. v. City and County of San Francisco, 217 F. Supp. 3d 1066, 1072-73 (N.D. Cal 2016) (cleaned up).
6 HomeAway.com v. City of Santa Monica, 918 F. 3d 676, 682 (9th Cir. 2019) (cleaned up).
7 Erie Ins. v. Amazon, 925 F. 3d 135, 139-40 (4th Cir. 2019).
8 501 (citing Musser v. Vilsmeier Auction, 522 Pa. 367, 562 A.2d 279 (1989)).
9 Oberdorf v. Amazon, 295 F. Supp. 3d 496, 502-503 (M.D. Pa. 2017).
10 Oberdorf v. Amazon. com, Inc., 930 F. 3d 136, 152-53 (3d. Cir. 2019).