“You have to choose: are you a platform or a publisher?”
It’s the question that makes us pull out our hair and roll our eyes. It’s the question that makes us want to shout from the rooftops “IT DOESN’T MATTER. YOU DON’T HAVE TO CHOOSE”
We’ll say it plainly here: there is no legal significance to labeling an online service a “platform” as opposed to a “publisher.” Yes. That’s right. There is no legal significance to labeling an online service a “platform.” Nor does the law treat online services differently based on their ideological “neutrality” or lack thereof.
There is no common law or statutory significance to the word “platform.” It is not found in Section 230 at all.
Some of the “You’re a platform!” mania is likely the fault of the companies themselves. Some have used the word “platform” to distinguish themselves, who primarily published user-generated content, from those who primarily published their own content, and/or actively edit and curate the content of others. They self-identified as “platforms” mostly to justify what was perceived as their hands-off approach to content moderation, particularly with respect to decisions not to remove hateful and harassing speech from their sites.
It’s fair to call out the big social media companies for holding themselves out as purely passive conduits (which is what some seem to mean when they call themselves “platforms”) when they actually moderate a ton of user content every day, and pretty much always have. Our work on the Santa Clara Principles reflects the human rights implications of content moderation, even though we support the First Amendment right of intermediaries to curate their sites.
But as a legal cudgel against perceived political bias, which is how the “admit it you’re a publisher not a platform” screed has most frequently been used, it is a meaningless distinction.
When politicians like Sen. Ted Cruz demand that Twitter identify itself as either a “publisher” or a “platform,” they usually make this false distinction in the entirely erroneous context of 47 U.S.C. § 230, the provision of U.S. law that grants broad immunity from liability to online intermediaries when such liability would be based on the speech of others. Rather than enshrine some significance between online “platforms” and “publishers,” Section 230 intentionally nullifies any distinction that might have existed. Contrary to popular misconception, immunity is not a reward for intermediaries that choose the path of total neutrality (whatever that means); nor did Congress enact Section 230 with an expectation that Internet services were or would become completely neutral. Section 230 explicitly grants immunity to all intermediaries, both the “neutral” and the proudly biased. It treats them exactly the same, and does so on purpose.
That’s a feature of Section 230, not a bug.
So online services did not self-identify as “platforms” to mythically gain Section 230 protection—they had that already.
Unlike “publisher” (more on that below), there is no common law or statutory significance to the word “platform.” It is not found in Section 230 at all. The word “platform” doesn’t even appear in any published Section 230 judicial opinions until 2004, and there and in most subsequent cases, the court simply quoted the descriptive language from the parties’ briefs in which it was used mostly as a synonym for “website.” Starting around 2010, courts did start using the word “platform” to describe internet services through which users interacted, much like courts used the terms “portal” or “website” previously and thereafter.
Moreover, regardless of Section 230, it is completely common to be both a “publisher” and a “platform” at the same time—a publisher of your own content and a platform for the content of others. Newspapers have historically done this and continue to do so—a publisher of the articles they write themselves and a platform for the content they publish but did not write themselves—letters to the editor, wire service articles, advertisements, op-eds, etc. And online publications and websites continue to do so now, mixing their own content with that generated by users.
In fact, it is really difficult to find any online service close to the user end, that is services like social media and email clients with which the user directly and openly interacts, that is solely a conduit for user speech, without any speech of its own. One doesn’t really find pure conduits like this until quite deep in the infrastructure layer of the Internet—like ISPs, domain name services, content delivery networks (CDNs), and email servers. And even that at depth, takedowns are not uncommon.
The specious publisher-platform argument is also historically off-base. Historically, there is some legal distinction between “publishers” and more passive “distributors” of others’ speech, and “distributors” is perhaps what those who yearn for “neutral platforms” are referring to. But “distributors” was just a subcategory of “publishers” and both bore liability.
So, what is the legal difference between “publishers” and “distributors”?
One is always a “publisher” of their own words, the stuff they write and say themselves. That is completely uncontroversial. The controversy and confusion arise around republication liability, the idea that you are legally a “publisher” of all statements of others that you republish even if you accurately quote the original speaker and attribute the statement to them. So, if you accurately and directly quote someone in an article you have written, and the quoted statements defame someone, you can be liable for defamation for republishing those statements. This applies to any content in your publication that you did not write yourself, like letters to the editor, advertisements, outside editorial, wire service stories, etc. Legally, you are responsible for all of these statements as if they were your own creations.
This legal concept of republication liability is an old concept inherited from English common law. But it appears that up until 1824, accurate attribution was a full defense.
A subcategory of these “publishers” are “distributors.” Since at least 1837, republication liability has extended also to mere distributors of speech—the 1837 case Day v. Bream dealt with a courier who had delivered a box of libelous handbills—if it could be proved that they knew or should have known about the illegal or tortious content. This “distributor” liability was widely applied to newsstands, booksellers, and libraries. The American version of this knowledge-based “distributor” liability is commonly associated with the US Supreme Court’s 1959 decision in Smith v. California, which found that a bookseller could not be convicted of peddling obscene material unless it could be proven that the bookseller knew of the obscene contents of the book. Outside of criminal law, US courts imposed liability on distributors who simply should have known that they were distributing actionable content.
So “distributor liability” applied to those like booksellers, newsstands, and couriers who merely served as fairly passive conduits for others’ speech, and “publisher liability” applied to those who engaged with the other person’s speech in some way, whether by editing it, modifying it, affirmatively endorsing it, or including it as part of larger original reporting. For the former, group, the passive distributors, there could be no liability unless they knew, or should have known, of the libelous material. For the latter group, the publishers, they were treated the same as the original speakers they quoted.
Because one was treated a bit better if they were a passive distributor, the law actually disincentivized editing, curation, or reviewing content for any reason.
One of the primary purposes of Section 230 was to remove this disincentive and encourage online intermediaries to actively curate and edit their sites without being so penalized. Former Rep. Chris Cox, one of the co-authors of Section 230, recalls finding it “surpassingly stupid” that before Section 230, courts effectively disincentivized platforms from engaging in any speech moderation. And Congress recognized that even the notice-based liability that attached to distributors created the prospect of the “heckler’s veto,” whereby one who wants the speech censored tells the distributor about it and the distributor removes the speech without devoting any resources to investigating whether the objection had any merit. As we have written (PDF), notice-based liability systems are subject to great abuse and have serious human rights implications.
Congress resolved this problem by getting rid of republication liability altogether. That is the significance of the phrase “No provider or user of an interactive computer service shall be treated as the publisher or speaker of any information provided by another information content provider.” And the very first federal appellate court to interpret Section 230 made clear that Section 230 got rid of both republisher liability and its subset, distributor liability.
So the very purpose of section 230 was to eliminate any distinction between those who actively select, curate, and edit the speech before distributing it and those who are merely passive conduits for it. For the purpose of Section 230, the only relevant distinction is between an “interactive computer service” and an “information content provider.”
But what about “neutrality”? It’s a bit more confusing, but there has never been any requirement that publishers be politically neutral or otherwise speak, edit, or curate without such decisions reflecting their own beliefs. And “neutrality” certainly doesn’t mean “ideological neutrality”—indeed, the First Amendment protects the right of any speaker to express their ideological and political viewpoints , whether through their own speech, or by choosing to serve as an intermediary for the ideological speech of others—that is by actively curating the user-generated speech on their sites.
With respect to republication liability, a few courts recognize a “neutral reportage” privilege to republish statements made by reliable speakers about a public controversy. In that context “neutral” means the that the statement is reported without any implication that it is true, but simply reporting that the statement was made. But the neutral reportage privilege has mostly been rejected by U.S. courts.
With respect to Section 230, the Ninth Circuit, gave the provision of “neutral tools” to users as one example of engagement with user speech that would preserve Section 230 protection. That is, one did not become the creator of content by providing “neutral tools” to help a speaker speak. The court did not explain what tools are in fact “neutral” and there remains confusion as to exactly what “neutral tools” are. But it seems the court meant that the service did not materially contribute to the illegality of the speech. Merely having a viewpoint or taking sides in a controversy would thus not negate Section 230 protection.
Section 230 aside, both “publishers” and “distributors” are liable only when the speech they disseminate is independently illegal or tortious. The mere act of editing or curating your users’ speech is not actionable.
To the contrary, curation and editing is itself speech protected by the First Amendment, as the Supreme Court has held with respect to platforms ranging from newspapers to St. Patrick’s Day parades. And at the center of this constitutional protection is the right to express one’s political views through the curation of others’ speech. As the Supreme Court stated, “the expression of editorial opinion […] lies at the heart of First Amendment protection.”