As FCC Commissioner Jessica Rosenworcel said of the Internet, “It is our printing press.” Unfortunately, for First Amendment purposes, regulators and courts treat our modern printing presses — electronic media — very differently from the traditional ones. Therefore, there is persistent political and activist pressure on regulators to rule that Internet intermediaries — like social networks and search engines — are not engaging in constitutionally-protected speech.
Most controversial is the idea that, as content creators and curators, Internet service providers are speakers with First Amendment rights. The FCC’s 2015 Open Internet Order designates ISPs as common carriers and generally prohibits ISPs from blocking Internet content. The agency asserts outright that ISPs “are not speakers.” These Title II rules may be struck down on procedural grounds, but the First Amendment issues pose a significant threat to the new rules.
ISPs are Speakers
Courts and Congress, as explained below, have long recognized that ISPs possess editorial discretion. Extensive ISP filtering was much more common in the 1990s but still exists today. Take JNet and DNet. These ISPs block large portions of Internet content that may violate religious principles. They also block neutral services like gaming and video if the subscriber wishes. JNet offers several services, including DSL Internet access, and markets itself to religious Jews. It is server-based (not client-based) and offers several types of filters, including application-based blocking, blacklists, and whitelists. Similarly, DNet, targeted mostly to Christian families in the Carolinas, offers DSL and wireless server-based filtering of content like pornography and erotic material. A strict no-blocking rule on the “last mile” access connection, which most net neutrality proponents want enforced, would prohibit these types of services.
The extensive filtering these ISPs do is not as common as it once was, but they illustrate the reality that ISPs do have editorial discretion and they do exercise it. For that reason, scholars have pointed out for nearly a decade that any meaningful no-blocking rule compels speech from ISPs and other Internet platforms. ISPs would be forced to transmit content that they object to.
Therefore, the Title II rules, especially a no-blocking rule, may not survive a First Amendment challenge. Title II proponents frequently assert that ISPs are “dumb pipes” and must be prohibited from exercising any editorial control over what is transmitted from consumers. Net neutrality activists around the globe, for instance, want to prohibit Internet.org, a free app jointly produced by Facebook and wireless carriers, because it “blocks” access to content that is is not selected to be included within the package.
To avoid scrutiny from a court, the FCC will need to show that ISPs resemble common carriers like telephone companies and FedEx that — though transmitting speech — don’t have editorial discretion and have essentially no First Amendment rights. However, if ISPs instead resemble electronic media like cable TV companies or search engines that exercise editorial control over transmitted content, the Title II rules represent compelled speech and will receive significant court scrutiny.
Internet Platforms and Curation
A court holding that ISPs are speakers for First Amendment purposes is important because Internet-based, curated distribution is replacing traditional ways consumers accessed news and media — bookstores, newsstands, broadcast radio, cable TV. The nature of “publication” and “speakers” has changed rapidly in a short time because information is much more accessible in the Internet age. Clay Shirky notes that the traditional formula of “Filter, then publish,” has been replaced with “Publish, then filter.”
While the nature of media changed, many still want to regulate the intermediaries. Google’s algorithm is a common target for regulation. Politico recently published a piece calling for search engine regulation because of Google’s ostensible ability to sway close elections through opaque algorithm tweaks. Many online companies and media companies likewise want to regulate the order in which Search results appear, an effort gaining traction in Europe. Further, some academics and activists would like to extend neutrality rules to Twitter’s and Facebook’s curation of user streams.
Fortunately, those efforts would likely fail in the US because Internet intermediaries receive constitutional protection. As Eugene Volokh persuasively argues, “search engines are speakers” and regulations affecting Google’s algorithms must withstand First Amendment scrutiny. Law professor Jonathan Zittrain advises regulatory caution and notes that “content curators…have a First Amendment right to present their content as they see fit.” ISPs likewise have an existing right (seldom exercised) to curate and filter content. As net neutrality supporter Harold Feld says, if the FCC doesn’t classify ISPs as common carriers, “nothing requires your ISP to deliver [Internet content]. If Comcast decides I am evil and blocks [my website], they can do it.”
Courts call this ability to filter, anachronistically, “editorial discretion.” In this new media world of content abundance, “curation” better represents what defines protected speech because most of the actual messages transmitted originate from third parties. Certainly, print organizations have substantial editorial control over what is published, but radio and TV is less controlled and newer media platforms display a wide range of editorial controls. Many of the most important modern speakers in this “publish, then filter” environment are curators. New media, like aggregators, cable companies, search engines, and ISPs, often use an intentional, semi-automated, iterative process to decide what content to omit and what to transmit.
Constitutional protection of curators means regulators likely cannot force a Christian cable operator to carry Cinemax. Likewise, Apple can continue to block apps containing Confederate flags or nudity in the App Store. Even though ISP filtering of content may be a blunter tool, like a WISP operated by Christians or Jews that blocks websites for its religious users, or like Internet.org that transmits only select Internet content, the First Amendment protects that ability to tailor content.
The Dumb Pipes Myth
Readers of law review articles from net neutrality advocates and of the Open Internet Order are left with the false impression that ISPs are passive transmitters — dumb pipes — that never block content. Title II supporters have to maintain this facade because if they suggest that ISPs exercise editorial control, the no-blocking rules trigger First Amendment scrutiny. Fortunately for First Amendment purists, there is sufficient evidence of ISPs filtering content to raise serious questions about the constitutionality of the Open Internet Order. Internet service providers can and do engage in filtering, and some curate content in ways that are much more intentional than First Amendment-protected speech activity like cable TV distribution and Google search results.
There are, as mentioned, small ISPs like JNet and Clean Internet that provide Internet access marketed to religious users. In addition to its wireline service, DNet is also a wireless ISP and advertises that they filter pornography and other content. Sprint, one of the Big Four national wireless carriers, last year offered social media plans and parental controls, including whitelists and blacklists, through its Virgin Mobile subsidiary.
I’ve posed this question to Title II advocates several times — Aren’t religious ISPs engaged in First Amendment-protected speech? To date, none have denied it and seem content to pretend server-based filtering by wired and wireless access providers doesn’t exist. Susan Crawford, for instance, wrote a law review piece about ISPs’ purported lack of First Amendment protection. She noted, correctly, that if ISPs were engaged in editorial decisions about filtering content, that would pose a problem for the FCC regulations. However, she abruptly discontinued a Twitter back-and-forth with me when I pointed out she omitted instances of religious ISPs exercising the editorial discretion she fears.
It’s not just smaller ISPs that filter. AT&T, for instance, like nearly every major ISP and Web company, reserves the right in its acceptable use policy to pull down content “that is determined by AT&T to be obscene, indecent, hateful, malicious, racist, defamatory, fraudulent, libelous, treasonous, excessively violent or promoting the use of violence or otherwise harmful to others.” This is not idle language. Groups like the Anti-Defamation League, for instance, knowing that they have enforceable acceptable use policies actively lobby and persuade ISPs and Web companies to remove content from anti-Semites and groups like the KKK.
Title II proponents like to point out that the large ISPs engage in relatively little of the curation and filtering that I’ve described. Therefore, they reason, ISPs are common carriers. I’m not persuaded. That distinction appears immaterial considering the FCC made no such distinction in its Open Internet Order. Further, large ISP reluctance to offer, say, family-friendly Internet packages is entirely predictable considering the FCC has for a decade chilled that exercise of free speech through ham-fisted attempts at net neutrality enforcement, merger conditions, and punitive fines.
Congress Intended to Encourage ISPs to Filter Content with Section 230
The FCC faces another obstacle to its conclusory determination that ISPs are not speakers. The 2015 Open Internet Order largely adopts net neutrality proponents’ First Amendment arguments and concludes that ISPs “serve as mere conduits for the messages of others, not as agents exercising editorial discretion.” However, Section 230 of the Communications Decency Act has a Good Samaritan provision that suggests ISPs are speakers, a view that several courts have endorsed.
For better or for worse, Section 230 makes ISPs and other Internet platforms that primarily rely on third-party content “super speakers.” Internet platforms can exercise editorial control and get all the benefits of being a speaker, like First Amendment protection, yet are immunized from many of the burdens, like liability for distributing online defamation and libel.
Why did Congress take this dramatic step in the mid-1990s? Quite simply, the drafters wanted to encourage ISPs to continue to block offensive content online. At that time, ISPs and bulletin board operators like Prodigy marketed themselves as family friendly and (inconsistently) blocked content that Prodigy administrators judged to be in bad taste. Because of this editorial discretion, in 1995 Prodigy faced costly liability in a defamation suit for defamatory statements one of its users posted, in a case called Stratton Oakmont v. Prodigy. Because of Stratton Oakmont and a few similar cases, Internet intermediaries faced two undesirable options to avoid liability for transmitted content:
- become conduits and exercise no editorial control — thereby leaving even offensive content online; or
- constantly police Internet content and take down all questionable material.
Congress disliked both options and quickly responded with Section 230 protections in 1996 to protect Internet-based distributors from becoming mere conduits. The statute protects “interactive computer services,” which includes, “specifically a service or system that provides access to the Internet….” Congress, therefore, preserved ISPs’ editorial role in cleaning up the Internet.
Several court cases recognize that ISPs and other Internet platforms exercise editorial discretion and are not mere conduits. As the district court said in the 1998 case Blumenthal v. Drudge, 230’s protections serve “as an incentive to Internet service providers to self-police the Internet for obscenity and other offensive material.” Similarly, the 9th Circuit Court of Appeals noted in the 2008 Roommate.com case that through 230, “Congress sought to [allow ISPs] to perform some editing on user-generated content….” A 10th Circuit decision states that “Congress clearly enacted § 230 to forbid the imposition of publisher liability on a service provider for the exercise of its editorial and self-regulatory functions.” Finally, as the Fourth Circuit said in Zeran v. America Online, Section 230 “forbids the imposition of publisher liability on a service provider for the exercise of its editorial and self-regulatory functions.”
At the very minimum, we have several courts saying that ISPs exercise editorial discretion — directly contradicting the assertions by Title II proponents and the FCC. Further, not only can ISPs filter content, it is a practice that Congress wished to encourage. The Title II rules chill those editorial functions, and that is a problem for the FCC.
Finally, while not a Section 230 case, a majority of the Supreme Court has tacitly endorsed Congress’ view that ISPs are speakers that can and should serve as curators of online content. In Ashcroft v. ACLU, the Court cited ISP filtering favorably as an alternative to unconstitutional provisions of the Child Online Protection Act. Taking these cases together, it’s unlikely a court will sustain the net neutrality advocate view that ISPs are not speakers and cannot engage in filtering.
The 2015 Open Internet Order and its Content-Neutrality Problem
Section 230 poses an additional problem for the FCC. Most courts construe Section 230(c)(2) broadly in terms of what services are protected and the kinds of liability providers are immunized from. A broad reading may protect ISPs from FCC regulations that restrain ISP filtering abilities.
The FCC’s hastily-written Title II order seems to recognize this but reveals some internal tensions in an effort to allow some Section 230-type ISP filtering. For example, while the Order says ISPs are not First Amendment speakers, the Order appears to accept the premise that Section 230 restrains agency action. Specifically, the Order likely allows ISPs to block offensive content and offer family-friendly packages because the FCC permits, in paragraph 220, ISPs to block “traffic that is unwanted by end users.” The FCC cites to the portions of the 2010 Open Internet Order that allowed Internet access packages that block, specifically, pornographic content. In those referenced portions of the 2010 Order, the FCC cites Section 230 and expressly states that the agency will not impose liability for good-faith actions by ISPs to restrict harassing and offensive content.
This exception to the no-blocking rule, if it is indeed an exception, puts the FCC in a bind when defending its no-blocking rule against First Amendment challenges. As an initial matter, this exception that allows ISPs to actively block the content described in Section 230 suggests FCC acknowledgement that ISPs are exercising editorial control and engaged in protected speech.
Further, if the FCC interprets this “unwanted traffic” exception narrowly, the agency would allow ISPs to block only lewd, harassing, and violent material (that is, the material specifically mentioned in Section 230), but would penalize ISPs for blocking, say, political, religious, and entertainment content. Such a distinction means that the Title II rules are not content-neutral regulations and therefore likely to be struck down on First Amendment grounds. In the words of the Supreme Court, courts will “apply the most exacting scrutiny to regulations that…impose differential burdens upon speech because of its content.” If, on the other hand, the FCC interprets “unwanted traffic” broadly in a content-neutral manner, the rules allow widespread filtering of Internet content, undermine the FCC’s assertion that ISPs are not speakers, and eviscerate the entire purpose of the Title II rules.
Before concluding, one aside is probably necessary: This position — that ISPs have a constitutional right to filter Internet packages for consumers — is often misrepresented and maligned by net neutrality advocates. To be clear, that ISPs or other Internet platforms have a First Amendment right to select which online content to transmit does not mean anticompetitive blocking or throttling of, say, Netflix is permissible. That is illegal. It likewise does not mean ISPs can promise subscribers access to Internet content and subsequently block that content. That is also illegal. This is a narrower claim: The First Amendment protects ISPs’ prerogative to transparently offer kid-friendly and other filtered Internet packages like Internet.org.
The future of the media is Internet-based and content is increasingly curated. It’s important that regulators don’t deprive ISPs of their First Amendment rights because ISPs represent the canary in the coalmine. First Amendment protection has generally been a one-way ratchet that’s expanded free speech protections for several decades and has been used successfully to fight regulations that affect speech. A setback for the First Amendment would encourage media access activists to seek more regulation for new media. The “gatekeeper” theory used to justify regulation of ISPs is just the warmed-over “scarcity” rationale for the Fairness Doctrine and other restraints on speech. The calls to regulate Twitter, Facebook, and Google algorithms and Internet television will grow even louder if courts accept the FCC’s First Amendment analysis in the Open Internet Order. Fortunately, the Open Internet Order suffers from several deficiencies and there is a good chance courts will again remind regulators to keep their hands off the Internet.
Article was originally posted on Plain Text on September 3.
The Technology Liberation Front is the tech policy blog dedicated to keeping politicians' hands off the 'net and everything else related to technology.