(A substantially similar version of this appeared first on the blog of my employer, Public Knowledge)
It’s hard to believe Trump issued this stupid Executive Order a mere week ago. Even by the standards of insanity known as the Trump Administration, the last week has reached heights of insanity that make a full frontal assault on the First Amendment with anything less than tear gas and tanks seem trivial. Nevertheless, given the vital importance social media have played in publicizing the murders of George Floyd, Ahmed Arbery, and too many others, how social media have broadcast police brutality against peaceful protesters to be broadcast live around the world from countless locations, and how social media has allowed organizers to to coordinate with one another, we need to remember how vitally important it is to protect these means of communication from being cowed and coopted by the President and others with power. At the same time, the way others have used social media to spread misinformation and promote violence highlights that we have very real problems of content moderation we need to address.
In both cases, Trump’s naked effort to use his authority to threaten social media companies so they will dance to his tune undermines everything good about social media while doing nothing to address any of its serious problems. So even though (as I have written previously) I don’t think the FCC has the authority to do what Trump wants (and as I write below, i don’t think the FTC does either), it doesn’t make this Executive Order (EO) something harmless we can ignore. Below, I explain what the EO basically instructs federal agencies to do, what happens next, and what people can do about it.
More below . . . .
Bashing social media for supposed liberal bias has become pretty standard fare for some conservative pundits and politicians. This remains true despite zero evidence of any kind of bias by social media companies against conservative content or Republican politicians. In fairness, Democratic political leaders have made similar accusations. For example, Speaker Nancy Pelosi (D-CA) has accused Facebook of “schmoozing” the Trump Administration and agreeing to “mislead the American people” in exchange for regulatory favors. Former Vice President Joe Biden has accused Mark Zuckerberg of knowingly collaborating with Russia and North Korea to rig the election in 2016 for Donald Trump and to undermine Biden’s own candidacy in the 2019-2020 campaign. It must be noted, however, that the only evidence we have of social media bias against an identity or ideological group, thus far, pertains to bias against African Americans.
Despite this slew of studies, President Trump in particular continues to press these claims loudly and frequently. It revs up the base and deters needed content moderation from the platforms. This tactic, according to the Wall Street Journal, has worked to scare Facebook off of its more effective proposals to combat disinformation since they would impact popular conservative speakers. This tactic tries to conflate “conservative speech” with protecting the civil rights of users on social media platforms. Last August, the President threatened to go after the shield from liability for third-party content online services enjoy under Section 230 (47 U.S.C. §230) as punishment for allegedly discriminating against conservatives and promoting liberals and Democrats. (You can read a boatload about Section 230 and what Public Knowledge thinks about it here.) Trump’s idea, as reported at the time, involved the Federal Communications Commission (FCC) issuing rules interpreting Section 230 as imposing an obligation of “viewpoint neutrality” (i.e., no fact checking the President or his allies), so that any sort of fact checking or other “censorship” based on viewpoint would cost a company its liability shield.
At the time, I wrote a fairly lengthy blog post explaining why the FCC has no authority to do any such thing. Additionally, as I noted then, it would require the FCC to completely reverse everything it said in the “Restoring Internet Freedom Order” about how Section 230 prohibits the FCC from imposing any kind of neutrality rules and – in a crowning irony – re-impose net neutrality on ISPs because (since ISPs are no longer Title II telecommunications providers) they are “interactive service providers” as defined by Section 230. Fortunately for everyone, Trump soon lost interest and the issue went away.
But now, thanks to Twitter marking two inaccurate and misleading Trump tweets about mail-in voting as inaccurate and misleading, Trump is apparently serious about it again. Last week he signed an “Executive Order on Preventing Online Censorship” which – when you strip away the rhetoric, basically requires three substantive things.
- The National Telecommunications Information Administration (NTIA) will file a Petition for Rulemaking with the FCC to ask them to make rules interpreting Section 230 as being dependent on “political neutrality” and maintaining themselves as a “public forum” without “viewpoint discrimination.”
- The Administration will forward the 15,000 complaints from conservatives it collected since August 2019 to the Federal Trade Commission (FTC) to see if it should pursue any sort of enforcement action under Section 5 of the FTC Act (15 U.S.C. 45) for “deceptively” claiming to be neutral (see Prager v. Google, citing Youtube’s terms of service) when secretly discriminating against conservatives.
- The Order directs all federal agencies with advertising budgets that purchase advertising on social media platforms to report to the Office of Management and Budget within 30 days which platforms engage in “viewpoint discrimination” or other “deceptive” practices and to withhold federal advertising dollars from these platforms.
Needless to say, this is precisely the worst case scenario the writers of the First Amendment had in mind – the President using the power of his office to punish those who criticize him and advantage his political faction.
At the outset, what Trump seems to want is illegal. It’s unconstitutional, since it seeks to regulate companies based on their point of view. The interpretation of Section 230 that the EO puts forward has no support in the text of the statute, the legislative history, or any court decision. Briefly, the “good faith” requirement in 230(c)(2), which permits online platforms to remove any content they find “objectionable, whether or not such material is constitutionally protected” does not impose any requirement of viewpoint or political neutrality. Online platforms of all sizes — from a small neighborhood bulletin board to the largest social media sites — are free to use their own judgment in deciding what posts to moderate or take down. (Even if a platform was found to have acted in “bad faith” for a particular content removal, it would simply mean there was no liability shield for that particular removal.) Further, 230(c)(2) places no conditions whatsoever on 230(c)(1), which unconditionally states that an online service cannot be held liable as a publisher or speaker of third-party content that stays up on the platform. There may be areas where we can debate whether Section 230 should be reformed. But the President can’t do it through an Executive Order (EO), and agencies that have no jurisdiction over the statute to begin with can’t change the statute through administrative action.
But others have covered these substantive issues in more depth. I want to cover the somewhat more geeky question of what happens next, how it works, and – the most asked question in America – who can sue about what.
How Does the FCC Process Work?
The FCC is an independent agency. Congress deliberately insulated it from direct control of the President in order to prevent the President from using it against political opponents and to ensure representation by both political parties. As events keep showing, this turned out to be a pretty smart move.
So the President can’t just order the FCC to do something. By law (47 U.S.C. §902(b)(2)(L)(i)), the NTIA acts as the interface between the Executive Branch and the FCC. Under the FCC’s rules (47 C.F.R. §1.401), anyone can file a Petition for Rulemaking or Petition for Declaratory Ruling (47 C.F.R. §1.2). The EO orders NTIA to file a Petition for Declaratory Ruling by Wednesday, July 29 Monday, July 27 (sorry, I miscalculated) asking the FCC to clarify that doing stuff contrary to the “purpose” of Section 230 as defined in the EO makes you lose the liability shield promised by Section 230. Since I’m primarily interested in process here, I won’t go into the rather tortuous legal reasoning for how this is supposed to work. Suffice it to say the NTIA (after consulting with the Attorney General) is supposed to file a Petition for the FCC to interpret Section 230 the way Trump wants and issue rules to enforce this interpretation.
This part is perfectly legal. Nothing stops the NTIA from filing a Petition with the FCC asking it to do anything arguably within the FCC’s jurisdiction. If the President orders NTIA to file a Petition with the FCC declaring that every hour broadcasters will include with their station identification “Trump is the best President ever” or that your cell phone start up screen has to say “MAGA – Republicans Rule Democrats Drool” they can Petition for that. Because anyone can file a Petition and those things are arguably in the FCC’s jurisdiction. It does not matter if they are silly, or if the FCC might not have the authority to actually grant the Petition.
Next, the FCC puts the Petition on Public Notice. It’s important to know that this is usually a semi-automatic process and doesn’t imply anything one way or the other about what the FCC intends to do. Generally, the Petition is referred to the relevant Bureau (not sure what it would be in this case), given a docket number (usually starting with RM, for “Rule Making,” but sometimes given a standard docket number if it is a matter of broad interest), and then released as part of the FCC’s daily business as a Public Notice (PN), usually saying “the following Petition for Rulemaking was filed by ___ on [date]. Comments may be filed by [30 days from release of PN], reply comments [15 days later]. These numbers can be changed. But again, this is pretty standard.
It is true that on rare occasions the Commission will intervene and simply sit on a Petition. It is also true that the Commission can move right to a Notice of Proposed Rulemaking in response to the Petition and issue that. But these are relatively rare and I would not expect those here on such a complicated and controversial issue. Also, as a political matter (more on that below), it works better to go the usual route and issue a standard PN relatively soon after NTIA files.
Public Comment Cycle. Once the FCC issues the PN and assigns a docket number, interested parties/members of the public can file comments with the FCC through the Electronic Comment Filing System (ECFS). This, of course, is where the fun really begins. In theory, parties file on the comment date and reply date. But in reality, the FCC designates these proceedings as “permit but disclose,” meaning you can file with the FCC and meet with FCC officials right up until the FCC is ready to make a decision.
What happens after public comment cycle closes? Most Petitions for Rulemaking never go any further than getting a docket number. Eventually, the FCC closes them out when they do a review of stale proceedings. But if the FCC wants to do something, it will issue a Declaratory Ruling that Section 230 works the way the President wants (or, if the FCC does this in a new Administration, a ruling that it does not have jurisdiction to issue rules under Section 230). Assuming they find for the President, the FCC will simultaneously issue either (a) rules implementing the Declaratory Ruling based on the Petition; or, (b) a Notice of Proposed Rulemaking about such rules.
Immediate Impact Will Not Produce New Rules, But Will Potentially Have Huge Political Impact and Stifle Platform Efforts to Combat Election Disinformation.
If the NTIA and the FCC stick to the anticipated schedule, the NTIA will file on July 29. The FCC will drop its PN sometime in the middle or end of August. This will make Comments due in mid-late September, with replies due in early-mid October. The observant among you will note that this just happens to coincide with when election season will be going full throttle and Trump will be looking for ways to rally his base.
While this is purely speculation, I would suggest this is not coincidence. It is an excellent way to intimidate social media companies from doing anything that might stoke backlash from conservatives that would translate into mass public comments (aka “working the ref”). Of course, the President and his allies will organize his base to file comments anyway, but it is still likely to make companies very reluctant to fact check or otherwise “censor” the President and Republican candidates at peak election time; i.e. just when the dirty tricks and disinformation are most likely to ramp up and when we most need social media companies to step up their game fighting election-related misinformation and disinformation.
Besides, after spending four years packing the bench with loyal judges, it’s worth a shot.
How Does the FTC Process Work?
The EO also states that the White House will forward the complaints it received through the tech bias reporting tool it launched last year to the FTC to investigate whether social media companies have behaved in an “unfair or deceptive” manner under Section 5 of the Federal Trade Commission Act. Again, anyone can file a complaint with the FTC over anything, so this action in and of itself doesn’t do anything. The FTC will then consider the complaint, which is something of a black box. Most complaints to the FTC not only go nowhere, you never hear about them ever again. This is one of many reasons people have become critical of the FTC and accused it of being toothless on consumer protection.
Assuming the FTC investigates however, it has a number of hurdles to overcome before it can bring a complaint. First, as anyone who has ever tried to read the terms of service of a social media provider (or any other online service for that matter) knows, these things are super-lawyered to give the companies lots of cover. So even if they say up front, “We totally love free speech and open debate and create a forum where a 100 flowers bloom,” the next 30 pages of fine print explain how the company can still take down whatever it wants, censor whatever it wants, do whatever it wants, and you have zero right to expect anything based on these terms of service, have a nice day. The FTC can also try to show that any supposed “censorship” is “unfair” rather than deceptive, but the burden of showing something is unfair to consumers is fairly high due to Section 5(n) and the narrow way courts have interpreted it. Basically, unless you can show monetary damages, it has gotten very hard to prove to a court that a company’s actions are “unfair” to consumers. As a consumer advocate, I am not happy about this, but that is how things are unless Congress changes the law.
But even after all that, you still have Section 230 and its liability shield. While Section 230 does not provide a shield from any federal criminal law, it does shield providers from civil liability (assuming it applies in the first place). For the FTC to find social media providers liable for civil penalties for unfair or deceptive practices under Section 5, a court would need to accept the President’s interpretation in the Executive Order – an interpretation that courts have universally rejected. No court has found that the liability shield of Section (c)(1) and (c)(2) are interlinked and dependent on each other. To the contrary, every court that has considered the question has found these two provisions separate and independent grounds for liability immunity. Courts have also universally rejected the claim that any sort of filtering, editing, labeling, or fact checking of third party content voids the liability shield of Section (c)(2) by making the interactive service provider an author/creator of the content.
Finally, while I am personally a big fan of the public forum doctrine invoked by the EO, and I believe (as I wrote in The Case for the Digital Platform Act) that based on Pruneyard Shopping Center v. Robbins, Congress could pass laws protecting and enhancing the “public forum” nature of social media, but courts have rejected the idea that social media services are intrinsically a public forum. The Ninth Circuit considered and rejected this argument just a few months ago in Prager University v. Google. So once again, I have great difficulty seeing the FTC pursuing an enforcement action here, and if they did I can’t see it winning.
Once Again, It’s More About Working the Ref Than Winning the Case.
Given the long odds of the FTC bringing a case, let alone winning one, why bother? Once again, it doesn’t cost President Trump anything to try (although it costs the taxpayers a bundle). Furthermore, investigations can be used to harass companies. The FTC can depose company officers and ask all kinds of pointed questions about their supposed political biases, demand expensive discovery production, tie up lots of company lawyers. It can then just drop the matter, then start the whole thing over again the next time Trump wants to “send a message” to a social media provider (and the industry generally) to dance to his tune.
Sure, these are (mostly) big companies that can handle the cost (although there are a bunch of little guys out there like Dreamwidth who would get crushed fairly easily if Trump wanted to make an example of someone). But even for the big companies, no executive likes getting deposed and cross-examined. Investigations create regulatory overhang that can depress stock value. It is one more way of “working the ref” and making companies feel uncomfortable about taking any action that might anger Trump, powerful Republicans, or popular conservative pundits — especially during the lead up to the 2020 election.
Additionally, and just as important, prosecutions for political purposes undermine the legitimacy of genuine cases that agencies should bring. Even threats to use legal investigations as a tool of harassment undermine efforts by agencies to enforce the law when required. For example, when the Department of Justice brought an antitrust action to block the AT&T/TW merger, AT&T argued in court that the case was a political vendetta over CNN’s often critical coverage of the President. While the district court rejected this challenge, it placed a cloud over the legitimacy of the investigation. Even if the FTC never brings an action as a result of the EO, this potentially undermines the legitimacy of any future antitrust or Section 5 enforcement action against giant social media companies as being driven by vindictiveness rather than by merit, even though such enforcement actions may be sorely needed.
How Does the Business With Political Advertising Work?
The final piece of the EO may actually have some direct near term impact on social media companies’ bottom line. The Order basically says that the federal government won’t spend advertising dollars on social media platforms that Trump thinks engage in “viewpoint discrimination” or other “censorship” or “deceptive practices.”
Generally, the federal government has broad discretion on how it spends money as a purchaser of goods and services, and can use that power to promote policies and social goals. For example, federal contracts generally contain a provision prohibiting discrimination based on sexual orientation, even in states that permit discrimination based on sexual orientation. Additionally, on the surface, the idea makes sense. We would not want federal advertising dollars supporting services that support racism or promote violence, for example. Under Walker v. TX Division, Sons of the Confederacy, Inc. the government has discretion to control government speech (such as advertising) and can therefore, in theory at least, withhold it to promote specific policy goals.
But there are limits. The big one here is something called the “unconstitutional condition” doctrine. Basically, if the Constitution prohibits the government from directly limiting a constitutional right, the government cannot indirectly limit that right by making a benefit contingent on waiving that right. This gets particularly tricky in the speech area. For example, in Rumsfeld v. Forum for Academic and Institutional Rights, the Court found that it was okay for Congress to require any school that received federal funding to give equal access to military recruiters that it gave to any other employer and this did not constitute “forced speech.” But in Matal v. Tam, the Court found that it violated the First Amendment to condition grant of trademark protection on whether or not the term or image sought to be trademarked was offensive.
In this case, you have a pretty standard problem for the Trump Administration. They said the quiet part out loud. Even in the EO, you can’t get away from the very specific targeting of Twitter for flagging the President’s tweets. This makes it very hard to pretend that this is about promoting diversity of views and dissemination of news from diverse and antagonistic sources – “a government purpose of the highest order” – rather than the presidential revenge scenario the First Amendment tries to prevent. (Indeed, by name checking the companies, the President arguably creates a Bill of Attainder problem, but that is a digression for another time.)
So Who Gets to Sue to Stop It and When?
That’s an excellent question. The Center for Democracy and Technology has already sued to block the entire EO. CDT faces two main obstacles to getting a hearing on the merits: standing and ripeness. Standing is the legal doctrine that requires a party to have a “particularized interest” in the matter at issue distinct from the general interest of the public in seeing a law enforced or right protected, that is, a “real, non-hypothetical injury in fact, arguably redressable by an order from the court.”
CDT makes the argument that as subscribers and users of social media, they are impacted by the chilling effect of the Executive Order. While this impacted class is large, “subscribers of social media” is sufficiently different from the general public to probably qualify. The hardest challenge to overcome is that the impact to CDT is likely to be considered too “remote” and “speculative” under current standing doctrine. The Supreme Court has previously held that fear of impacts based on how government action will impact a third party are too speculative to confer standing. See Clapper v. Amnesty International USA (unwillingness to talk to foreign clients by phone for fear that intelligence agencies may monitor phone calls too speculative and remote for standing purposes). On the other hand, the courts are often more charitable to the potential “chilling effect” on free expression as an injury than in other cases. So while I would put the odds against CDT, it is hardly a slam dunk.
The companies themselves certainly have standing to sue, either directly or through their trade associations. Here, however, you run into the problem of ripeness. Under this doctrine, courts won’t hear a case that depends on something too hypothetical happening in the indefinite future. After all, the bad thing might never happen, and the court can’t really evaluate the circumstances based on something that might possibly happen but hasn’t actually happened yet.
Here, the actions that flow from the EO that potentially cause the harm are pretty far off and hypothetical because they are committed to agency discretion. The FCC may not put the NTIA Petition out for public comment, and even if it does it might either do nothing or deny it. The FTC may or may not choose to investigate. Even the advertising review is uncertain, since the agencies could conclude that all the social media services they purchase advertising from behave appropriately and in compliance with federal policies. This makes it very easy for a court to say: “sure, even if you have standing, nothing has really happened yet. Come back when agencies actually do something.”
But once again, we have this idea of “chilling effects” on speech. Courts recognize that when it comes to behavior, people don’t like to get arrested or have their advertising dollars or liability shields threatened. In such cases, the mere threat to take action may make the matter sufficiently ripe to seek a court order declaring the Administration’s actions unconstitutional. I normally would not give this much chance, but the egregious conduct by the Administration (and the enthusiastic cheerleading by at least one FCC Commissioner) make the “chilling effect”/intimidation argument stronger than usual.
Even if parties have to wait, however, they will certainly get their day in court if the agencies press forward. Participants in an FCC proceeding who can show standing can sue, and demonstrating injury for social media subscribers will be easier if there are rules in place governing how social media services must behave. The subjects of any FTC investigation and complaint will certainly have the right to defend themselves in court, and companies that suffer any loss of advertising dollars can challenge that decision in court under the unconstitutional condition doctrine.
Conclusion
On the one hand, we shouldn’t expect any immediate legal consequences from the President’s Executive Order. The primary short-term impacts will likely be political rather than legal. That doesn’t excuse what amounts to a blatant assault on the First Amendment for naked political advantage. Furthermore, although Twitter, at least appears at the moment to be willing to stay the course on flagging tweets that violate its policies, we should expect that social media companies will find it difficult to stand up to the pressure in the long run.
So what can concerned members of the public do? First, even supporters of Section 230 reform should make it clear that the President’s Executive Order is not a welcome way to move forward — as suggested by FCC Commissioner Carr — but a cynical hack of our legal and regulatory systems by the President of the United States, to the detriment of the First Amendment and the general legitimacy of the Rule of Law. We cannot forget that this Executive Order was brought on because the President was rightly called out for spreading voting misinformation. We must not overlook that point as it is crucial to how truly cynical and harmful what the President is attempting to do is, and how it strikes at fundamental rights of participation in our democracy. We cannot regard this as excusable, trivial, or harmless.
Second, assuming NTIA files a Petition and the FCC puts it on public notice, we must have widespread public pushback. Politicians and social media companies alike need to see that the American people are not dupes, nor willing to tolerate turning the Rule of Law into a protection racket for whoever holds the White House. Republicans as well as Democrats should fear what happens if we allow the President to brazenly weaponize the machinery of justice through jaded silence and cynical acquiescence.
Stay tuned . . .
