Topic: Telecom, Internet & Information Policy

Why Section 230 Is Unstable

Everyone interested in social media should read Jeff Kosseff’s The Twenty-Six Words that Created the Internet. It provides an excellent history of Section 230, the legal foundation of social media. That might sound boring, but Kosseff makes it work by mixing stories and analysis without vitiating either. I agree with Kosseff that, problems notwithstanding, the benefits of Section 230 have outweighed its costs.

Given that, I don’t look forward to future editions of his book since they may document the “fall” of Section 230. In other words, the law is likely to be amended to limit the protections offered internet platforms. We have already seen changes meant to combat sex trafficking. Section 230’s most serious persistent vulnerability, however, comes from a mismatch between its statutory language and the larger world the law inhabits.

Our political world is divided between (let’s call them) Progressives and Conservatives. Progressives see society as a struggle between designated oppressor and oppressed groups. No one – not government, not the tech companies – should be neutral between these groups; everyone should favor the oppressed. Since the speech of oppressors is causally essential to the harms experienced by the oppressed, government and the platforms should suppress that speech to help the oppressed. Remember, no one is neutral in this endless struggle.

As it happens, given the extent of Progressive cultural authority, the language of Section 230 favors the Progressive cause. Here’s an edited version of Section 230(c)(2) that clarifies this point:

No provider or user of an interactive computer service shall be held liable on account of any action voluntarily taken in good faith to restrict access to or availability of material that the provider or user considers to be…objectionable, whether or not such material is constitutionally protected…[emphasis added]

So if content moderators think speech offending the oppressed is “objectionable,” they can banish it from their service.

Professor Tim Wu Makes The Case Against Antitrust Policy

It is common to hear proponents of antitrust action against big tech firms talk up the potential for future harms to consumers from sustained dominance by Facebook, Google, and Amazon. 

In her influential “Amazon’s Antitrust Paradox,” lawyer Lina Khan argued that “the current market is not always a good indication of competitive harm” and that antitrust authorities should “ask what the future market will look like.” This sentiment was recently echoed by economist Jason Furman in a digital competition review for the UK government.

One of the best cases against such an approach was inadvertently delivered by long-time antitrust champion Professor Tim Wu at a Stigler Center conference on antitrust last week. While critiquing the consumer welfare standard approach in a debate with Tyler Cowen, Wu said:

everyone who is even vaguely honest as an economist will agree that dynamic costs matter more than static costs and dynamic benefits matter more than static benefits. But those are the hardest to measure, so we’ve gotten trapped in a world where the old joke about the economist and the street light has become the soul of the law.

Exactly. Antitrust policy can indeed tend to think of markets too statistically. But antitrust enforcers also have no special insight into the future of markets and available technologies, and hence the change in the balance of costs and benefits to consumers going forwards. Looking at a static market may well lead to mistaken conclusions. But it’s a complete leap of faith to presume that replacing static analysis (more accurate but incomplete) with dynamic analysis (supposedly comprehensive but wildly speculative) will deliver better outcomes for consumers.

The Federal Election Commission Is Bad Enough

Chris Hughes, a founder of Facebook, has proposed Congress create a new agency to “create guidelines for acceptable speech on social media.”

As Hughes notes, this proposal “may seem un-American.” That’s because it is. At the very least, Hughes’ plan contravenes the past fifty years of American constitutional jurisprudence, and the deeply held values that undergird it. Let’s examine his case for such a momentous change.

He notes that the First Amendment does not protect all speech. Child porn and stock fraud are not protected by the First Amendment. True threats as harassment are also illegal. Incitement to violence as understood by the courts can also be criminalized. All true, though more complex than he admits.

The fact that the courts have exempted some speech from First Amendment protection does not mean judges should create new categories of unprotected speech. Hughes needs to make a case for new exemptions from the First Amendment. He does not do so. Instead he calls for an agency to regulate online speech. But, barring drastic changes to First Amendment jurisprudence, his imagined agency would not have the authority to broadly regulate Americans’ speech online.

However, Hughes’ old firm, Facebook, can and does regulate speech that is protected from government restrictions. In particular, Facebook suppresses or marginalizes extreme speech (sometimes called “hate speech”), “fake news” about political questions, and the speech of foreign nationals.

Facebook is not covered by the First Amendment.  You can support or decry their decisions about such speech, but it would be “un-American” to say Facebook and other private companies do not have the power to regulate such speech on their platforms. And I might add that you can exit Facebook and speak in other forums, online and off. A federal speech agency would not be so easily avoided.

Hughes may think that Facebook is doing a poor job of regulation and that its efforts require the help of a new government agency (which would be subject to the First Amendment). But if we ended up with government speech codes imposed by private companies, the courts might well swing into action on the side of free speech. In that sense, the new agency would actually vitiate private efforts at content moderation. We might well end up with more of the online speech Hughes and other critics want to restrict.

In sum, Hughes’ agency is a bad idea in itself. It is unlikely to accomplish his goals. The agency might even weaken private efforts to limit some extreme speech. Of course, if judicial doctrines changed to accommodate new speech restrictions imposed by this new agency, America really would change for the worse. It is encouraging, however, how little support and how much criticism Hughes’ proposal has received from his fellow Progressives. (See the critiques by Ari Roberts and Daphne Keller linked here).   Conservatives should feel free to chime in.

Armslist and Bias against Conservatives Online

Last week, conservatives once again cried “bias” after Facebook banned a spate of popular fringe pundits and conspiracy theorists. Meanwhile, the week’s most important content moderation story went, for the most part, unnoticed. Had conservatives paid more attention to the Wisconsin Supreme Court’s ruling in Daniel v. Armslist, they might feel differently about the utility of platform intermediary liability protections like Section 230 of the Communications Decency Act, a bedrock indemnity that prevents internet platforms from liability for user behavior. While usually understood merely as a shield for social media firms, it guards a wide variety of services that utilize user-generated content, such as classified advertising or individual websites’ comments sections.

Armslist is essentially a digital classified ads section for guns. Daniel, the daughter of a shooting victim, sought to hold Armslist liable for the use of its platform by her mother’s murderer. Her suit alleged that certain features of Armslist’s site were negligently designed, without regard for how they might be used by persons prohibited from buying firearms. According to the complaint, Armslist should have anticipated that its lack of user registration requirements and the ability to search for in-state, background check-free sales would be misused by patrons prohibited from possessing firearms. In most states, it is perfectly legal to privately sell a firearm to your neighbor without utilizing the services of a federally licensed dealer. The imposition of broad liability on services that help to coordinate legal activities would burden and perhaps preclude Americans’ right to sell and buy firearms, a legal activity.

The suit against Armslist represents a growing trend of attempts to circumvent CDA 230’s protections by suing platforms for features that enable certain kinds of harmful user behavior, rather than simply suing over the behavior itself. Snapchat was recently sued for the creation of a speedometer filter by plaintiffs who were struck by a Snapchat user driving at over a hundred miles an hour (the driver was speeding in pursuit of a high speedometer reading on the Snapchat app, attempting to impress her friends with her foolishness). Thankfully, in both the Snapchat case and now in Daniel v. Armslist, judges have understood that “no matter how artfully pled”, these suits attempt to hold platforms responsible for user behavior and are therefore precluded by CDA 230.

Digital classified ads services and speedometers are neutral tools. Chief Justice Patience D. Roggensack writes in Armslist: “All of these features can be used for lawful purposes, so the CDA immunizes interactive computer service providers from liability when these neutral tools are used for unlawful purposes.” So long as a given feature can be used lawfully, service providers cannot be held liable for their unlawful use: doing so would unreasonably burden lawful users of the tool in question. Just because Uber can be used to summon a getaway vehicle after a heist does not render Uber a getaway car-hailing service. Had Armslist been decided differently, or if CDA 230’s protections were to be limited or eliminated, these tools would not be commercially viable. 

The conservative claims about social media bias may lead to legislative revisions to CDA 230. Those revisions could easily restrict current CDA 230 protections for businesses like Armslist. Indeed, some on the left would see removing such protections as a goal of revising CDA 230. The Armslist decision shows that the harm done to the Second Amendment would be real and permanent. Are the speculative gains of seizing control of Facebook’s content moderation really worth the risks to the Second Amendment? Won’t this be a case of unintended consequences of the sort conservatives used to warn us about so many, many years ago?

Apple v. Pepper: A Chip Off the Old Illinois Brick?

In yesterday’s decision in Apple v. Pepper, Justice Brett Kavanaugh joined the four liberal Justices to rule that class action lawyers can sue Apple on behalf of consumers who allegedly paid uncompetitively high prices for iPhone apps, even though the consumers bought the apps not from Apple itself but from third-party developers who were paying a commission to the tech giant. The majority rejected Apple’s defense under the so-called Illinois Brick doctrine, under which only direct purchasers of a good or service, but not purchasers further down the distribution chain, can sue over monopoly pricing (everyone agrees that current law empowers the developers themselves to sue Apple for alleged monopolistic behavior). Kavanaugh, on behalf of the majority, said Apple lost the benefit of the Illinois Brick defense when it inserted itself into the supply chain as a retailer through its Apple Store, thus making itself in practice an intermediary even if it was not itself the party deciding what to charge app buyers. 

The significance of yesterday’s ruling is probably not in its proximate consequences for the iPhone supply chain, which are still uncertain (the ruling allows the plaintiffs to proceed, but doesn’t mean they’ll win). As Justice Neil Gorsuch observed in dissent, the Court’s mini-rule is “pointless and easily evaded”: 

To evade the Court’s test, all Apple must do is amend its contracts. Instead of collecting payments for apps sold in the App Store and remitting the balance (less its commission) to developers, Apple can simply specify that consumers’ payments will flow the other way: directly to the developers, who will then remit commissions to Apple. No antitrust reason exists to treat these contractual arrangements differently, and doing so will only induce firms to abandon their preferred—and presumably more efficient—distribution arrangements in favor of less efficient ones, all so they might avoid an arbitrary legal rule.

The wider worry, as Gorsuch points out, is that the majority (significantly joined by Kavanaugh) did not merely resolve a technical puzzle about how the law’s language applies to an unusually designed supply chain, but seemed inclined along the way to adopt an ungenerous and narrow reading of Illinois Brick. And that is significant because in the overall scheme of antitrust law, Illinois Brick serves as a major check against runaway litigation (aside from its own logic, it restrains multiple and duplicative suits over the same behavior). For that reason, the antitrust plaintiff’s bar has long sought to knock down the defense. Yesterday’s outcome gives it a passing chip at most, but will bear watching as a harbinger. 

The Ruins of Old Tech Monopolies

Look on my works, ye Mighty, and despair!

Mar 2000: Palm Pilot IPOs at $53 billion

Sep 2006: “Everyone’s always asking me when Apple will come out with a cellphone. My answer is, ‘Probably never.’” – David Pogue (NYT)…

Jun 2007: iPhone released

Nov 2007: “Nokia: One Billion Customers—Can Anyone Catch the Cell Phone King?” (Forbes)

Geoffrey Manne and Alec Stapp at Truth on the Market have written a brief history of impregnable tech monopolies that were pregnable after all, in fields from personal computers to video distribution to social media. Sen. Elizabeth Warren and others are now arguing that the government should break up and closely regulate tech giants Google, Amazon, Facebook, and Apple “claiming they have too much power and represent a danger to our democracy.” Manne and Stapp offer examples of sector after sector in which what was seen as structural, inescapable tech monopoly turned out to be not so unassailable. Here’s music distribution: 

Dec 2003: “The subscription model of buying music is bankrupt. I think you could make available the Second Coming in a subscription model, and it might not be successful.” – Steve Jobs (Rolling Stone)

Apr 2006: Spotify founded

Jul 2009: “Apple’s iPhone and iPod Monopolies Must Go” (PC World)

Jun 2015: Apple Music announced

 They conclude by quoting two observations by Benedict Evans, a venture capitalist at Andreessen Horowitz, first on “why competition in tech is especially difficult to predict”:

IBM, Microsoft and Nokia were not beaten by companies doing what they did, but better. They were beaten by companies that moved the playing field and made their core competitive assets irrelevant. The same will apply to Facebook (and Google, Amazon and Apple).

And why “we will not be stuck with the current crop of tech giants forever”:

With each cycle in tech, companies find ways to build a moat and make a monopoly. Then people look at the moat and think it’s invulnerable. They’re generally right. IBM still dominates mainframes and Microsoft still dominates PC operating systems and productivity software. But… It’s not that someone works out how to cross the moat. It’s that the castle becomes irrelevant. IBM didn’t lose mainframes and Microsoft didn’t lose PC operating systems. Instead, those stopped being ways to dominate tech. PCs made IBM just another big tech company. Mobile and the web made Microsoft just another big tech company. This will happen to Google or Amazon as well. Unless you think tech progress is over and there’ll be no more cycles … It is deeply counter-intuitive to say ‘something we cannot predict is certain to happen’. But this is nonetheless what’s happened to overturn pretty much every tech monopoly so far.

Supreme Court Argument in the “Scandalous Trademarks” Case Wasn’t So Funny

FUCT may be f … Well, you get the idea. Two years ago in the Tam case, the Supreme Court struck down the Lanham Act’s “disparaging trademarks” provision, but the justices seem less likely to erase the “scandalous trademarks” prohibition now – at least as far as one can tell from this morning’s argument in Iancu v. Brunetti.

That’s because racial slurs and other offensive phrases necessarily have a viewpoint – on the basis of which the First Amendment doesn’t allow the government to discriminate – but swear words can be just a “mode of expression.” At the same time, the “scandalous” mark restriction is so broad that the government is asking the Court to either accept its benevolent assurances or narrow the statute. There were echoes of the government’s assurances not to prosecute certain kinds of speech in Citizens United there, and indeed the same deputy solicitor general, Malcolm Stewart, initially argued that case – leading to the Court’s setting it for re-argument and blowing up the relevant statute.

The most telling series of questions involved the regulation of bus advertising, and that example should indeed decide the case: you should be able to register trademarks that come short of obscenity (which is generally pictures or sentences, rather than single words), but both registered and unregistered trademarks are still properly subject to time, place, and manner restrictions. And that includes “limited public forums” like the sides of municipal buses, public park benches, and the like.

One other note: none of George Carlin’s seven dirty words were used during the argument (though the late comedian’s formulation was invoked several times). Stewart described FUCT’s “scandalous” homonym as the “past participle of paradigmatic profanity,” while Chief Justice John Roberts at one point asked (page 58): “I take it that the – a correct spelling of the vulgar word at the heart of the case, that can’t be trademarked, right?”

We’ll know by the end of June whether edgy marks are up Schitt’s Creek.

I previously wrote about the background of Brunetti – in which Cato filed its latest “funny” (more “vulgar”) brief.


Editor’s note: This post originally misstated the chief justice’s question.

Pages