Are Social Media Services “State Actors” or “Common Carriers”?

Yesterday, I did a videoconference with Prof. Eugene Volokh (UCLA Law) discussing if and how legislatures could regulate Internet services. Watch the video. Prof. Volokh and I usually agree on most things, but this time we may reach different conclusions. I think Prof. Volokh might be open to the idea that a legislature could declare social media to be a “common carrier” and impose associated must-carry obligations. In contrast, I’m a hard no, and I hate the question because it only matters if someone hopes to censor social media despite the First Amendment.

The notes I made preparing for the event (we didn’t get to all of these questions):

Q: What is Section 230?

Goldman Answer: 

Section 230 has three operative provisions: (1) Section 230(c)(1): websites aren’t liable for third-party content. (2) Section 230(c)(2)(A): no liability for filtering decisions. (3) Section 230(c)(2)(B): no liability for filtering instructions.

Section 230(c)(1) eliminates the moderator’s dilemma, the phenomenon where it may be better to not try to manage content if you are liable for any mistakes you make. (c)(1) allows Internet services to dial up/down their content moderation efforts. Services can do no content moderation, or extensive content moderation, and the outcome is the same (no liability).

Section 230 is a speech-enhancing statute. The First Amendment sets a floor on speech, not a ceiling, and legislatures can and do protect speech more than the First Amendment requires (other examples: reporter shield, defamation retraction, Consumer Review Fairness Act). The procedural benefits of enhancing speech via Section 230 include: quicker and cheaper to dismiss and Constitutional avoidance.

Q: What would happen to social media companies without Section 230?

Goldman Answer: In a post-230 world, Internet services would choose among three options:

1. Moderate content vigorously, because they would accept liability for any mistakes they made. But (1) moderation is costly, (2) being sued is costly, and (3) the damages from suit could be overwhelming. For the (few?) services that can profitably pursue this option, they would engage in massive overremovals and giving fewer people access to their publication tools. In other words, the Internet shrinks a lot

2. Do as little moderation as possible, with the hope that services can claim that they didn’t know, and shouldn’t have known, of any tortious or criminal material. This fits the Cubby v. CompuServe approach and possibly Smith v. California. However, (1) takedown notices will function as hecklers’ vetoes, (2) plaintiffs will argue the service “should have known” of problems despite the minimal moderation, and (3) the spammers and trolls will overrun the forum. The latter phenomenon will drive away legitimate users, and the service will collapse from the weight of an unmanageable pile of garbage (reminds me of WALL-E).

3. Exit the industry.

So what will a post-230 Internet look like?

  • UGC incumbents become entrenched because costs go up, which drives out competitors and discourages new entrants. At best, Google and Facebook may be the only surviving UGC players.
  • A broad shift away from UGC and towards professionally produced content. Professionally produced content has lower risks of liability, and the licensors might provide an indemnity. In that shift:
    • Less free content would be available (ads won’t pay enough to compensate the professionally produced content).
    • We’ll see a reduction in the diversity of available content.
    • Fewer publication venues available for us to speak.
    • Many content categories will cease to exist. Ex: consumer reviews, online marketplaces, YouTube how-to videos, Wikipedia.
  • What will still be around:
    • Paywalled gardens, similar to Quibi.
    • Twitter would still permit brands and celebrities to publish (and maybe Eugene, but not me).

Q: Are social media companies “state actors”?

Goldman Answer: My position has always been no. I have blogged the caselaw many times (see the list below). To me, the Halleck case seems to resolve this question emphatically:

  • “In operating the public access channels, MNN is a private actor, not a state actor, and MNN therefore is not subject to First Amendment constraints on its editorial discretion.”
  • “Expanding the state-action doctrine beyond its traditional boundaries would expand governmental control while restricting individual liberty and private enterprise.”
  • “when a private entity provides a forum for speech, the private entity is not ordinarily constrained by the First Amendment because the private entity is not a state actor. The private entity may thus exercise editorial discretion over the speech and speakers in the forum.”
  • “Providing some kind of forum for speech is not an activity that only governmental entities have traditionally performed. Therefore, a private entity who provides a forum for speech is not transformed by that fact alone into a state actor. After all, private property owners and private lessees often open their property for speech. Grocery stores put up community bulletin boards. Comedy clubs host open mic nights.”
  • “In short, merely hosting speech by others is not a traditional, exclusive public function and does not alone transform private entities into state actors subject to First Amendment constraints.”

The Prager U v. YouTube (9th Cir. 2020) ruling reinforces this:

  • “Despite YouTube’s ubiquity and its role as a public-facing platform, it remains a private forum, not a public forum subject to judicial scrutiny under the First Amendment… YouTube does not perform a public function by inviting public discourse on its property”
  • Regarding YouTube providing a public function: “the relevant function ‘must be both traditionally and exclusively governmental.’”

Q: A North Dakota bill would make a social media site with 1M+ users liable for damages if it “restricts, censors, or suppresses information that does not pertain to obscene, lewd, lascivious, filthy, excessively violent, harassing, or otherwise objectionable subject matter.” Could this work?

Goldman Answer: I object to the word “censorship” in this context. Only governments censor. Private entities exercise editorial control.

The North Dakota bill would conflict with Section 230 by imposing damages for social media services liable for removing third-party speech. See Backpage v. McKenna (W.D. Wash. 2012). The Washington statute required age verification requirements on certain third-party ads. The court said this conflicted with Section 230 because (1) it imposed liability for third-party content; and (2) creating disincentives to monitor third-party content conflicts with Congress’ goals in 230.

The North Dakota bill would also conflict with the First Amendment. Social media services have their own independent First Amendment rights to freedom of speech and press. If social media services qualify for constitutional protections as publishers, the answer is an easy no. See Miami Herald v. Tornillo (1974), which struck down a Florida right-to-reply statute (i.e., a must-carry law). The court said:

the Florida statute fails to clear the barriers of the First Amendment because of its intrusion into the function of editors. A newspaper is more than a passive receptacle or conduit for news, comment, and advertising. The choice of material to go into a new paper, and the decisions made as to limitations on the size and content of the paper, and treatment of public issues and public official — whether fair or unfair — constitute the exercise of editorial control and judgment. It has yet to be demonstrated how governmental regulation of this crucial process can be exercised consistent with First Amendment guarantees of a free press as they have evolved to this time.

Note that the Tornillo court rejected the newspapers’ must-carry obligation despite the fact that (as the court explained in the opinion) newspapers were functionally monopolists in their municipalities.

Q: Could Congress declare internet infrastructure services such as Amazon Web Services to be common carriers, thereby requiring them to provide services to all without discrimination? 

Goldman Answer: The question implicates how we regulate different layers of the telecom stack. Some parts of the stack may be utilities. Others aren’t. Many “Internet infrastructure providers” do not resemble the situations that are normally considered common carriers, such as natural monopolies or other circumstances where it would be socially wasteful to make duplicate physical investments.

Also, unlike most “utilities,” Internet infrastructure providers routinely have the technical ability to do more than just turn their service on/off. Compare domain name registrars. They have only 1 remedy option (service on/off), with the added disadvantage that registrants can’t second-source registration with another provider (if service is terminated, they need to quickly find a replacement). Yet, registrars aren’t subject to any mandatory neutrality/must-carry law. Indeed, governments routinely deputize them to control their customers’ substantive behavior (see, e.g., how registrars were drafted into the opioids fight). [Note: because they are much lower in the stack, domain name registrars are a better candidate for common carrier treatment than web hosts].

The First Amendment does not permit many Internet services to be legislatively analogized to broadcasters. As Reno v. ACLU indicated, Red Lion, Turner, and Sable “provide no basis for qualifying the level of First Amendment scrutiny that should be applied to the Internet.”

Also, there is little justification for mandating carriage/non-discrimination when the infrastructure niches are not a natural monopoly, because multiple competitors allow customers to second-source. Parler’s real problem is that it sole-sourced its entire business to AWS–an own-goal.

Q: Do search engines qualify for Section 230 and First Amendment protection?

Goldman Answer (from my Internet Law casebook):

Google is protected by the First Amendment’s free speech and free press clauses. Thus, any regulatory mandate that Google include or exclude information in its search index is almost certainly unconstitutional. See, e.g., Search King, Inc. v. Google Technology, Inc., 2003 WL 21464568 (W.D. Okla. 2003); Langdon v. Google, Inc., 474 F. Supp. 2d 622 (D. Del. 2007); Zhang v., Inc., 10 F. Supp. 3d 433 (S.D.N.Y. 2014); Google, Inc. v. Hood, 96 F. Supp. 3d 584 (S.D. Miss. 2015) (vacated on other grounds); e-ventures Worldwide v. Google, Inc., 2017 WL 2210029 (M.D. Fla. 2017); Eugene Volokh & Donald M. Falk, First Amendment Protection For Search Engine Search Results, April 20, 2012; see also Martin v. Hearst Corporation, 777 F.3d 546 (2d Cir. 2015) (publication cannot be obligated to remove article about an expunged arrest).

Furthermore, Section 230 (both (c)(1) and (c)(2)) statutorily immunize search engines for their indexing decisions, including their refusal to de-index content (even if that content is tortious). See, e.g., Maughan v. Google Technology, Inc., 143 Cal. App. 4th 1242 (Cal. App. Ct. 2006); Murawski v. Pataki, 514 F. Supp. 2d 577 (S.D.N.Y. 2007); Shah v. MyLife.Com, Inc., 2012 WL 4863696 (D. Or. 2012); Merritt v. Lexis Nexis, 2012 WL 6725882 (E.D. Mich. 2012); Nieman v. Versuslaw, Inc., 2012 WL 3201931 (C.D. Ill. 2012); Getachew v. Google, Inc., 491 Fed. Appx. 923 (10th Cir. 2012); Mmubango v. Google, Inc., 2013 WL 664231 (E.D. Pa. 2013); O’Kroley v. Fastcase Inc., 831 F.3d 352 (6th Cir. 2016); Fakhrian v. Google Inc., 2016 WL 1650705 (Cal. App. Ct. 2016); Despot v. Baltimore Life Insurance Co., 2016 WL 4148085 (W.D. Pa. 2016); Manchanda v. Google, Inc., 2016 WL 6806250 (S.D.N.Y. 2016); Mosha v. Yandex Inc., 2019 WL 5595037 (S.D.N.Y. 2019); see also Yeager v. Innovus Pharmaceuticals, Inc., 2019 WL 447743 n.6 (N.D. Ill. 2019) (“no ‘right to be forgotten’ exists under United States law”).

Selected Related Posts About State Action Claims