Tag: Moderator’s Dilemma

It’s the Algorithms, Stupid! – Understanding the Section 230 Debate

By Chris Ricchetti | 27 January 2021

Our country is being torn apart by the polarizing effects of tech-enabled echo chambers… Strategic curation of the content presented to users is the very definition of what it means to be a speaker or a publisher… Both amplifying and constraining the reach of any user-generated content algorithmically should be illegal.

PREFACE – Up-Front Disclosure
PART I – Background
• What is Section 230
• Common Misconceptions
• Section 230 In Context
• Provisions of Section 230
• Modifications Since 1996
PART II – The Current Debate
• Political “Neutrality”
• Truth and Lies
• Complete Repeal
• Mixed Metaphors
• Barr’s Two Cents
• Getting Warmer
• Right Tool, Right Job
• It’s the Algorithms, Stupid!
AFTERTHOUGHTS – Lingering Legal Ambiguities
END NOTES

Updated 13 February 2021: To monitor the latest proposed and pending legislation related to Section 230, check the Disruptive Competition Project.

Center for Humane Technology Logo

To learn more about the effects of technology on human well-being, visit the Center for Humane Technology.

PREFACE – Up-Front Disclosure↑

This article provides a straightforward and relatively “objective” review of Section 230, the circumstances that gave rise to it, the intent behind it, the intended and unintended consequences of its enactment, judicial interpretations, and the robust public debate about potential changes to the law—none of which address the most serious and urgent tech-enabled threat to democracy: the reckless misuse of algorithmic technology.

Also included—at no charge—are the author’s point of view and analysis, which can be summarized as follows:

  1. There are three broad categories of problematic, user-generated content: (1) content that is clearly criminal, (2) content that is “objectionable” but not clearly criminal, and (3) content that is false, distorted, misleading, etc., but not clearly criminal. One (1) is the only category that government is constitutionally empowered to restrict. This does not make the other two any less destructive or absolve Americans and our government from seeking ways to lessen their detrimental effects.
  2. General protection from civil liability for third-party content must be the default position for internet platforms that host user-generated content—unless they knowingly facilitate or are negligent in moderating clear criminality. Section 230 protects commerce and innovation across the entire internet. It does not only benefit “Big Tech.” [1]
  3. However, the existing protections conferred under Subsection (c)(1) are too broad. Section 230 should not, as courts have frequently held that it does, protect platforms that knowingly facilitate clearly criminal content or activity, that are willfully blind, or that are negligent in moderating clear criminality. Such platforms should not be immune from liability and should be shut down.
  4. Lawmakers should define in detail the minimum required standards and practices for moderating clear criminality, including penalties for non-compliance (i.e., something other than the loss of Section 230 protections across the offender’s entire business).
  5. To the extent that overly broad interpretations of Section 230(c)(1) may continue to impede the government and injured parties from bringing civil actions in connection with criminal conduct, clarifying language should be added to Subsection (c)(1), without disturbing its vitally important broad civil liability protections. Obviously, the law should never be applied as a shield from culpability for any underlying criminal conduct.
  6. Lawmakers should not attempt to regulate the moderation of content that is “objectionable” but not clearly criminal (category “2” above). The largest platforms do this reasonably well, despite innumerable (and inevitable) errors, omissions, and inconsistencies.
  7. The tech-enabled amplification of lies, distortions, and misinformation (category “3” above) on a global scale is—by far—the most serious and destructive consequence of Big Tech’s ubiquity in American life.
  8. The use of algorithms to auto-curate user-generated content (including shared content not authored by the sharing user) should be categorically prohibited as a strict-liability tort. Offenders should be excluded from all liability protections under Section 230(c)(1) and be deemed publishers of all third-party content posted to their platforms.
  9. Withdrawing the liability protections granted under Section 230 is not the appropriate remedy for correcting perceived injustices related to discrimination on the basis of politics or the Freedom of Speech. Civil rights law is the proper forum in which to fight those battles, for those who are inclined to fight them.
  10. Withdrawing the liability protections granted under Section 230 is not the appropriate remedy for correcting anticompetitive markets. Antitrust law is the proper forum in which to fight those battles, for those who are inclined to fight them.
  11. Keeping children physically, emotionally, and morally safe and secure should always be of paramount importance in policymaking. There is no higher purpose under the social contract on which democracy is based. However, whenever social conservatives purport to do anything in the name of “protecting children,” there is cause for skepticism and scrutiny. Evangelical and Catholic leaders have a decades-long track record of appealing to the universally shared value of protecting children to justify legislation and law enforcement actions that were, in fact, intended to chill or to criminalize constitutionally protected adult speech. In retrospect, most of these efforts have failed to withstand judicial scrutiny and have been partially or fully struck down by the courts—including, in more than one instance, by the Supreme Court of the United States. Consequently, the producers of some of the most extreme and harmful content have either been punished very lightly or not punished at all. Society would be better served by the enactment of more narrowly tailored statutes that provide muscular protections for kids without advancing a broader, moralistic “hidden agenda.”

PART I – Background↑

The current debate about Section 230, part of the Communications Decency Act of 1996, is taking place at a time when Americans are deeply polarized. Technology has brought to light and greatly amplified our differences to such an extent that there are now, quite literally, two “Americas”—each with its own values, priorities, “facts,” and perceptions of reality. That Big Tech companies have become “too big” and “too powerful” is among the rarest of declarative statements that both Americas deem to be “true,” albeit for different reasons.

What is Section 230?↑

Title V, Section 509 of the Communications Decency Act of 1996,[2] entitled “Online Family Empowerment,” amends the Communications Act of 1934 by adding Section 230 at the end of Title II.

The heading of Section 230 reads, “Protection for Private Blocking and Screening of Offensive Material.”

In short, Section 230 encourages tech companies to censor content that is “obscene, lewd, lascivious, filthy, excessively violent, harassing, or otherwise objectionable” and shields them from civil liability when they do. In other words, content creators—including users who post comments—may not sue tech companies for labeling or removing their content in “good faith.” (Establishing “bad faith” is neither inexpensive, nor easy.)

Most of the 1996 Act was struck down by the Supreme Court[3] subsequent to its enactment, but the part of the Act commonly referred to as “Section 230” was deemed constitutional and remains in effect.

Common Misconceptions↑

Before examining the current debate about Section 230, let us look at some of the common legal misconceptions—among both left and right leaning Americans—that are clouding the debate.

The First Amendment to the Constitution of the United States prohibits the government from infringing upon Americans’ Right to Freedom of Speech, except when doing so serves a clear and compelling public interest. The First Amendment does not restrict citizens from taking action to limit, suppress, distort, or otherwise interfere with each other’s speech. There are many laws that do protect Americans from each other, but the First Amendment is not one of them. Social media platforms are private spaces, and no one has a constitutional “right” to freedom of speech on social media. In fact, the First Amendment protects the rights of social media platforms to moderate their private spaces without government interference. Their moderation decisions will ultimately be judged in the marketplace.

There are other important features of American civic life that many Americans have either forgotten or may never have learned. Many Millennial and Gen-Z Americans, by no fault of their own, were raised with a set of values from which they could easily and understandably infer certain inaccuracies about American law, particularly as it relates to the freedom of speech and discrimination. Consider the following examples:

Speech that is commonly referred to as “hate speech”—as immoral, destructive, and despicable as it may be—is not categorically illegal. More precisely, it is illegal only when it is direct, personal, and can be shown to be “truly threatening” or reasonably likely to incite others to immediate, illegal action, especially violence. As in all areas of the law, there are many nuances and an extensive body of case law in which the contours of these general principles have been worked out and applied to a wide variety of facts and circumstances. Consequently, the latitude to spew hatred is not as wide as it may seem. But there is a great deal of highly offensive speech, including some that many would consider to be “triggering,” that is nonetheless legal and protected under the First Amendment. While the subset of all hate speech that is prohibited by law is narrower than many Americans realize, Section 230 does enable tech companies, if they choose, to censor the complementary subset of all hate speech that is hateful but not illegal.

Discrimination—as immoral, destructive, and despicable as it may be—is also not categorically illegal. More precisely, it is illegal only in certain contexts and only when it affects a legally protected class of persons. The laws restricting discrimination are somewhat different in the areas of employment, housing, education, and public accommodations. The types of discrimination that are always illegal include discrimination on the basis of “race, color, religion, sex, or national origin.” In some parts of the country, private clubs and religious organizations are free to discriminate—even on the basis of these otherwise unlawful forms of discrimination. However, if a business owner has a personal distaste for rich people, or “latte liberals,” or people who eat meat, or doctors who perform abortions, or dog owners who don’t clean up after their dogs, or alumni of Michigan State, it is altogether legal for her to deny service for any of these reasons—provided that the reason given for the denial of service is not a “cover” for some other form of discrimination that is actually illegal. In June 2018, former White House Press Secretary, Sarah Huckabee Sanders, was ejected from the Red Hen restaurant in Lexington, Virginia because of her association with President Donald Trump. This discrimination was legal, and Ms. Sanders had no legal recourse. Likewise, social media platforms and other internet spaces are fully within their rights to deny service or to censor anyone on the basis of politics. And they are under no obligation to exercise discriminatory bias on a consistent basis.[4]

Those who have suggested (perhaps not without merit) that social media platforms of a certain size are de facto public spaces could pursue expanding the statutory meaning of “public accommodations” to include large internet platforms, and they could pursue adding “political views” to the list of protected traits[5] under the Civil Rights Act of 1964. Some countries already prohibit discrimination on the basis of politics. But Section 230 was not intended to deal with issues of discrimination and is not the proper legal tool to correct any perceived injustice related to discrimination.

Section 230 is also not about monopoly power. To the extent that the enormity, economies of scale, reach, and other market advantages of Big Tech companies may constitute nearly insurmountable barriers to entry and stifle competition, antitrust litigation is the proper legal tool to correct these anticompetitive market dynamics. Threats to withdraw Section 230 liability protections as means of “punishing” Big Tech for their success are misguided and will not change their oligarchical market positions or lessen their market dominance.

Lastly, it should be self-evident that Big Tech is wholly uninterested in functioning as arbiters of truth. We know this because they resisted public pressure to involve themselves in “truth moderation” and politics for a long time and, until recently, had made few attempts to censor even the most egregious lies and oppressive political content from despicable actors (e.g., foreign dictators). Worse yet, some smaller players have done exactly nothing upon learning that their platforms were facilitating horrific crimes. Big Tech is interested in one thing only: maximizing user engagement. Big Tech simply wants users on their social media platforms consuming content—pretty much any content—for as long as possible, every day, because that is what drives their revenue. From their perspective, robust content moderation is an enormously expensive and risky distraction from making money.

Big Tech is not in collusion with the “deep state,” or “antifa,” or George Soros, or the “squad,” or the “Illuminati,” or “a cabal of cannibalistic democrat celebrity pedophiles” to thwart democracy and establish a one-world communist government in fulfillment of Biblical prophecy. To the extent that “shadowbanning” was ever a “thing,” the practice was an early effort by Big Tech to quietly extract themselves from the irreconcilable demands of the Two Americas. Their moral failing has been their intentional amplification of polarizing falsehoods—not their reluctant efforts to censor lies, distortions, and actual criminal content in response to a sustained public outcry and threats from lawmakers of both parties to cut them off at the knees. As with Corporate America generally, Big Tech’s obsession with maximizing shareholder value without regard for any other moral precepts defines its depravity, not “censorship.”

Section 230 In Context↑

When Section 230 was enacted in 1996, public access to and commercial use of the internet were new, as was the graphical user interface. Section 230 was intended to address two specific—and closely related—issues of concern at the time.

First, before Section 230 was enacted, some courts held[6] that if internet companies restricted access to any third-party content, they would then become liable for all third-party content accessed via their services—to the same extent as if they had been the actual creators of the content. Few internet companies would have survived if they could have been sued for every instance of racist, sexist, discriminatory, libelous, obscene, or otherwise offensive content that users might encounter. The clear legal incentive before Section 230 was for internet companies to do no moderation at all. Many Americans wanted internet companies to take a more active role in “policing” and restricting access to offensive content. Section 230 was intended to correct the perceived problem of under-moderation.

Second, in the early years of the public internet, many Americans were concerned about the proliferation of pornography online (perhaps they were on to something) and they wanted to ensure that children would not have access to or inadvertently land on adult content. As the title of the legislation (Communications Decency Act of 1996) suggests, preventing minors from encountering adult content was the overarching and primary purpose of Section 230 and of the Act as a whole. In fact, the part of the Act that establishes Section 230 is entitled “Online Family Empowerment.” Specifically, the legislation was intended to enable internet access software companies—such as CompuServe and America Online—to deploy technologies that would empower parents to block adult websites and other content not appropriate for children, without being deemed publishers of and without assuming potential liability for all the content on all the websites across the entire internet that users accessed via their services.

The technology environment in 1996 was vastly different from what it is today. Users accessed the internet via the existing land-based telephone lines. There were no internet service providers (ISPs) apart from the land-based phone carriers.[7] There were no social media platforms. Online communities—called “bulletin boards”—were much smaller and tended to be populated by a small subculture of “brainiacs” and “geeks.” Most Americans were not even aware that they existed. There was almost no (non-adult) entertainment online. There was no streaming video—neither prerecorded nor live (content had to be downloaded, then viewed or played locally). There were no lifestyle or affinity websites, limited news, little commerce, and few platforms (other than the aforementioned “bulletin boards”) where users could post comments or engage in open dialogue. There were no mobile devices capable of accessing the internet or any of its content. All of this developed rapidly in the years following 1996, but none of it was known or contemplated at the time Section 230 was enacted.

Provisions of Section 230↑

Section 230 is both clear and concise.[8]

Subsections (a) and (b) of Section 230 state that the internet is super cool and that it is the policy of the United States to promote its ongoing development.

The substance of the legislation is found at Subsection (c), which reads as follows:

(c) Protection for “Good Samaritan” blocking and screening of offensive material

(1) Treatment of publisher or speaker

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.

(2) Civil liability

No provider or user of an interactive computer service shall be held liable on account of—

(A) any action voluntarily taken in good faith to restrict access to or availability of material that the provider or user considers to be obscene, lewd, lascivious, filthy, excessively violent, harassing, or otherwise objectionable, whether or not such material is constitutionally protected; or

(B) any action taken to enable or make available to information content providers or others the technical means to restrict access to material described in paragraph (1).[9]

Subsection (d) simply states that internet companies should notify users of any parental control features available on their platforms, and Subsection (e) confirms that Section 230 is not intended to affect the interpretation or enforcement of certain existing laws.

In short, Section 230 resolved the “Moderator’s Dilemma” (see footnote 6). Subsection (c)(1) says, in effect, “There is no need to censor anything and everything that could possibly be offensive to somebody somewhere, because no one can sue you anyway (i.e., don’t over-moderate).” The intent here was to keep the internet a place for vibrant and open discussion. Subsection (c)(2) encourages moderation where it truly matters by saying, in effect, “On the other hand, we don’t want you to completely ignore the third-party content on your platform either (i.e., don’t under-moderate). These are the types of content that we’d like you to make a good-faith effort to moderate…” The intent here was to make the internet safe for kids. The language of Section 230 enables, indeed, encourages, platforms to restrict user access to objectionable content without the risk of being deemed the creators of any content created by a third party.

Modifications Since 1996↑

In response to the Supreme Court’s decision to strike down all but Section 230 of the Communications Decency Act of 1996, Congress passed, and President Clinton signed into law, a more narrowly tailored effort to modify the liability protections of Section 230 called the Child Online Protection Act (1998). This legislation was also overturned by the High Court, on the grounds that it was still too overly broad, vague, and burdensome and was likely to chill legal speech between adults.

The Digital Millennium Copyright Act (1998), without directly modifying the language of Section 230, created new requirements for internet companies to respond to infringement complaints related to copyrighted content posted on their platforms, and opened them to potential liability for third-party content if they fail to comply with the “safe-harbor” procedures for processing infringement complaints set forth in the Act.

While Section 230 explicitly states that the legislation was not intended to affect criminal law,[10] it has nonetheless been cited by courts as the basis for shielding internet companies and their responsible agents from civil actions in cases where they had been knowing enablers of, or the principal actors in, criminal activity. In response, Congress passed, and President Trump signed into law, the Allow States and Victims to Fight Online Sex Trafficking Act of 2017 (FOSTA-SESTA). The Act clarifies that Section 230 does not prohibit the enforcement of Federal or state criminal or civil law when internet companies “knowingly assist, facilitate or support sex trafficking,” thus creating the first explicit statutory exception to the liability protections afforded by Section 230.

It appears to some that forcing sex trafficking offline may have had the unintended effect of making it more difficult for law enforcement to prosecute sex trafficking crimes and may also be making sex workers less safe.[11] Others have claimed that the Act stifles innovation in that it creates an oversized risk for tech entrepreneurs who could inadvertently run afoul of the law with devastating consequences. Add these to the laundry list of concerns that should be considered when evaluating proposals for further legislative action.

In the early days of the public internet, before anyone could conceive of the advances in technology that would enable its vast commercial potential or the development of a “digital divide,” broadband (i.e., internet-access infrastructure) was treated as an “information service” for regulatory purposes. This designation limited the powers of the Federal Communications Commission (FCC) to regulate broadband.

In February 2015, in response to growing public demands to preserve the neutrality and openness that made the internet great, the Obama FCC issued the Open Internet Order, that reclassified broadband as a “telecommunications” service under Title II of the Communications Act of 1934, subjecting it to the more stringent regulatory regime applicable to “common carriers” (e.g., phone companies). Among other consequences, the FCC was then able to ban internet service providers (ISPs) from “picking internet winners and losers” via practices such as blocking, throttling, and paid prioritization. Additionally, ISPs would now have a duty to work toward universal access to more uniform internet access services at more uniform prices. Ironically, ISPs would not now need the protections of Section 230 because their status as common carriers came with its own protections from civil and criminal liability. Internet actors higher on the “stack” than ISPs, including websites, and computer and mobile applications, remained very much reliant upon Section 230 protections.

In 2018, the Trump FCC issued the Restoring Internet Freedom Order (you can’t make this stuff up) restoring ISPs to their original, lightly regulated status as “information services,” thus defeating the Obama Administration’s noble efforts to maintain net neutrality.

In May 2020, former President Trump signed an executive order purported to strip Twitter, Facebook, and Google of Section 230 liability protections. Like most of Trump’s reactive, petty, and entirely uninformed initiatives, it went nowhere. It did, however, dovetail with a more substantive examination by the Barr Justice Department of all things related to Section 230, already under way, that culminated in proposed legislation being sent to Congress by Attorney General Barr (see below).

PART II – The Current Debate↑

With all this in mind, let us now turn our attention to the current debate about the content moderation practices of Big Tech and the arguments being made for amending or repealing Section 230.

Political “Neutrality”↑

We’ll start with an easy one—the Ending Support for Internet Censorship Act. In June 2020, insurrectionist and Dunning-Kruger poster boy, Senator Josh Hawley (R-MO) proposed requiring tech companies to “prove” their “political neutrality” (i.e., that their moderation practices reflect no “discernable viewpoint”) to the Federal Trade Commission every two years as a condition for Section 230 immunity. Really, Josh? Government control of private speech? Flag on the play. First Amendment violation. 6-year penalty. Next!

Senator Hawley’s proposed legislation and his messaging around it are cynical, misleading and in bad faith. As a Yale Law graduate, he well understands that his proposed “solution” to the “problem” of Big Tech “censorship” is unconstitutional on its face. Senator Hawley simply happens to be, for the moment, one of the more visible performers carrying out the standard Republican modus operandi: flagrantly lying to their own constituents, for whom they have neither respect, nor concern. Instead of leading by telling the truth about issues and proposing workable solutions, they opportunistically seize upon every grievance du jour, and milk it dry to augment their own power. None but a small handful of Republicans stand for any discernible consistent ideology beyond their own self-interest. Hawley has no regard for the freedom of speech as a shared value, and no concern for the insidious injustices of discrimination, except in this one instance where he perceives it to be affecting him. He is plenty well prepared by education and experience to lead his constituents back to reality and to come up with meaningful proposals to address their real concerns. And he knows that Section 230 is not the right tool for the job.

As previously stated, it may well be that large social media platforms should be regarded as public accommodations[12] under the law, and it may well be that discrimination on the basis of politics should be legally prohibited. However, each would represent a fundamental change to American law that would have implications far beyond the scope of Section 230. If these more foundational changes were to be enacted, cleaning up the language of Section 230 to bring it into conformity would then become little more than a ministerial function.

Truth and Lies↑

Concerns about “fact checking” and “truth arbitration” (i.e., moderating intentionally misleading or factually incorrect content) should be kept separate and distinct from concerns about discrimination on the basis of politics and the freedom to express dissenting opinions (see previous section). They are two different things altogether.

In the current deeply polarized environment, solving the problem of pervasive misinformation is surely the more challenging of the two. No one trusts Big Tech to do it, and everyone seems to believe that the slightest intrusion by government would take us down a steep slippery slope to authoritarianism—even though most Americans were willing to accept significant limits on their civil rights following the tragic events of September 11, 2001, ostensibly in service of the public good.

Beyond the excruciating question as to who should be charged with truth arbitration, lies the equally daunting challenging of defining the line between alternative points of view and “alternative facts.”

The difficulty of these challenges should not preclude trying to solve them—at least to some extent that would make a meaningful difference. In a pluralistic, democratic society, the scope of publicly accepted truth arbitration would have to be narrow and limited to that which can be transparently and empirically verified, then confirmed from a multitude of diverse perspectives. As philosophers have observed for thousands of years, objectivity is elusive. But facts are not relative. Our deeply divided house will not stand unless we can achieve some degree of reconciliation between the binary “realities” of the Two Americas.

Many participants on all sides of the current debate do not seem to realize that all of this is beyond the scope of anything that Section 230 was ever intended to address. A Section 230 makeover will not deliver us from the existential crisis that we face.

Complete Repeal↑

Both President Biden and former President Trump have advocated for the complete repeal of Section 230. Speaker Pelosi has called it a “gift” that “could be removed.” This is a terrible, “surgery-with-a-chainsaw” remedy that would have numerous far-reaching consequences, both good and bad.

While Section 230 has unquestioningly been misused to shield some very “bad hombres” from accountability for serious and harmful crimes (this can be fixed), its effect on the growth and development of the internet has been largely positive. Section 230 has even been credited with singularly shaping the development of the internet as we know it.[13]

Section 230 generally protects tech companies—as well as startups and hobbyist bloggers who host comments—from liability for user-generated content. This essential protection is a fundamental necessity for doing business on the internet, not a special privilege. We all benefit from innovation and competition among the providers of tech-enabled services—from home delivery to entertainment to local transportation and travel to the availability of vast information and educational resources to matchmaking. Removing this liability protection would, in the long run, harm the “little guy” more than it would harm Big Tech.

As understandably frustrated as so many are with Big Tech and its enormous influence on American life, the total repeal of Section 230 is not the proper remedy. To maintain an online environment that is vibrant, open, dynamic, and convenient, there must be reasonable limits on the financial risks associated with developing new applications and launching new ventures. Hobbyists publishing recipes and cooking videos on a private URL should not be held liable for offensive content that passersby may decide to post on a whim. Startups that are fortunate enough to experience sudden geometric growth should not be forced to prioritize and allocate disproportionate resources to human moderation of user-generated content during a make-it-or-break-it growth phase. Imposing this burden on “small tech” by law would create yet another barrier to entry and further shield Big Tech from meaningful competition.

Mixed Metaphors↑

A bill introduced in the Senate during the second session of the 116th Congress by Senators Graham (R-SC) and Blumenthal (D-CT) called the Eliminating Abusive and Rampant Neglect of Interactive Technologies Act of 2020 (EARN IT) would establish a National Commission on Online Child Sexual Exploitation Prevention, under the control of the Attorney General, to recommend best practices that platforms may (or may not) choose to implement to prevent and respond to instances of child sexual abuse on their platforms. The original bill would also have carved out a second statutory exception to Section 230 liability protections for tech companies that fail to comply with the Commission’s recommended best practices or demonstrate to the Commission that they have in place their own reasonable practices for preventing child sexual exploitation on their platforms. That provision was not included in the substitute version of the bill reported to the Judiciary Committee in July 2020.

The original version of the bill came under fire for an entirely different reason. The original bill would also have guaranteed “legal access” by law enforcement to any digital message, a provision that critics claimed would effectively ban the use of encryption technology in the United States—yet another debate for which Section 230 is not the proper forum. Not only was the new requirement regarding encryption removed from the substitute version of the bill, the substitute bill also expressly protects platforms from both civil and criminal liability for child sexual exploitation based solely on their use of encryption technology—even when it impedes law enforcement from gathering evidence.

Barr’s Two Cents↑

In June 2020, Attorney General William Barr sent proposed legislation to Congress that would condition immunity from liability for moderation decisions upon compliance with the moderation parameters set forth in his amended Subsection (c)(2). Specifically, moderation decisions would be shielded from liability only when platforms have an “objectively reasonable belief” that the moderated content is “obscene, lewd, lascivious, filthy, excessively violent, promoting terrorism or violent extremism, harassing, promoting self-harm, or unlawful, whether or not such material is constitutionally protected,” and only when they moderate in “good faith,” a requirement that Barr’s proposed legislation defines in detail.[14] Any other moderation decisions would not be shielded from liability under Section 230, although the general exemption from civil liability for unmoderated non-criminal content would remain in effect. The proposal is clearly intended to discourage “censorship” of political content. It might discourage some platforms from moderating most or all “category 2” (objectionable but not illegal) content.

Second, the Barr proposal would remove liability protections in circumstances where platforms purposefully “promote, solicit, or facilitate” third-party content or activity in violation of any Federal criminal law (called the “Bad Samaritan Carve-Out”).[15] Third, just as FOSTA-SESTA created an exception for sex trafficking content (see above), the Barr proposal would create three additional, specific exceptions—for child abuse and exploitation, terrorism, and cyberstalking content. Fourth, the Barr proposal confirms that the Federal and state governments could bring civil actions against platforms in connection with any failure to remove criminally unlawful content under one of the statutory exceptions to Section 230.

Barr’s proposed legislation is not unreasonable. Platforms of all sizes would continue to enjoy broad protections under Section 230, except when they cannot provide a “reasonable factual basis” for their moderation decisions. The addition of a “Bad Samaritan Carve-Out” would certainly be positive to the extent that it reinforced the ability of law enforcement to prosecute actual criminality that causes serious harm.

The “good faith” standards set forth in the Barr proposal are quite superior to those included in the Limiting Section 230 Immunity to Good Samaritans Act introduced at about the same time by Senators Josh Hawley (R-MO), Marco Rubio (R-FL), Mike Braun (R-IN), Tom Cotton (R-AR) and Kelly Loeffler (R-GA).[16] The Senators introduced the bill in deference to former President Trump’s Executive Order aimed at exacting revenge on Twitter, Facebook, and Google (see above). Undoubtedly, the Senators’ principled effort was further inspired by a series of lawsuits brought against social media platforms by aggrieved “influencers,” such as far-right sensation Laura Loomer, among many others, that had failed to yield any relief for the plaintiffs from the pervasive tyranny of anti-conservative bias.

As previously mentioned, under current law, social media platforms are private spaces, and discrimination on the basis of politics is not prohibited. Moreover, efforts by the government to regulate the moderation of legal content are likely unconstitutional under the First Amendment. It may well be that exceptionally large social media platforms have become de facto public spaces and should be treated as such for public accommodations and civil rights purposes. But Section 230 is not the proper forum for resolving such foundational and far-reaching issues.

Getting Warmer↑

The Platform Accountability and Consumer Transparency Act (PACT Act), introduced in June 2020 by Senators Brian Schatz (D-HI) and John Thune (R-SD), appears to be a sincere effort at meaningful reform that is uncomplicated by hidden motives. First, the Act would establish new transparency and reporting requirements that most major platforms already practice. Platforms would have to publish detailed acceptable use policies, detailed moderation standards, and quarterly statistics tracking their moderation decisions. All good. Second, platforms would have to provide users with mechanisms for reporting acceptable use violations and potentially illegal content, and for appealing moderation decisions.[17] Also good. Third, the Act would establish standards for processing, acting upon and responding to content-related complaints. Under certain circumstances, if platforms failed to comply with their own policies or applied them inconsistently, they could be subject to Federal Trade Commission enforcement actions for “unfair or deceptive acts and practices.” Points for creativity! Fourth, Section 230 protections would not apply in the case of certain violations, such as failing to remove illegal content within 24 hours of being notified by courts or users.

The proposed legislation would seem to incentivize platforms to adopt lenient acceptable use policies and to over-moderate, making it easier for malicious users to submit frivolous complaints to silence other users. Also, the Act would provide exceptions for smaller internet businesses—generally a fantastic idea! However, specifically with respect to criminality that causes serious harm, the smaller players probably commit some of the most egregious offenses.

Overall, the PACT Act would be a step in the right direction—a scalpel, rather than a chain saw. The Act deals primarily with the moderation of content that is actually illegal, as well as transparency in moderation practices generally, and it does not seek to misapply Section 230 to solve unrelated popular frustrations.

Right Tool, Right Job↑

The Digital Millennium Copyright Act (1998) and the Stop Internet Sexual Exploitation Act—a bill intended to rein in platforms that facilitate distribution of non-consensual porn (AKA “revenge porn”) introduced in December 2020 by Senators Jeff Merkley (D-OR) and Ben Sasse (R-NE)[18]—are two examples of targeted legislation that creates appropriate and effective regulation to solve specific problems related to third-party content while leaving the broad general liability protections of Section 230 untouched.

It’s the Algorithms, Stupid!↑

If any of the advancements in computer technology since 1996 rise to the level of a paradigm shift, it is the capacity to compile gargantuan amounts of data about individual users, married with sophisticated algorithms that drive an unending stream of custom-curated content into the mind of each individual technology user. This hyper-automated bias confirmation complex plays upon the neural plasticity of the human brain, continuously reinforcing of our tribal beliefs. The colossal rift between the Two Americas is being hardwired into our brains at warp speed. Our country is being torn apart by the polarizing effects of tech-enabled echo chambers.

It is not their good-faith efforts to fact-check and remove egregious lies—and certainly not their efforts to prevent clear criminality—that renders the extension of Section 230 immunity to large social media platforms indefensible. Their destructive use of algorithms is what harms the public good! Big Tech’s reckless use of machine learning to maximize user engagement time by exploiting fear and outrage, for the ultimate purpose of monetizing users, in willful blindness to the devastating consequences, should be categorically prohibited as a strict-liability tort.[19]

Strategic curation of the content presented to users is the very definition of what it means to be a speaker or a publisher.

The use of algorithms for some purposes should be acceptable. Deploying user data and algorithms for the purpose of serving micro-targeted advertising content is a defensible practice. Advertising content is reviewed and approved by platforms prior to publication and, therefore, is not third-party content and, therefore, is not shielded from liability under Section 230 anyway. The utilization of algorithms as moderation tools (e.g., auto-removal of content containing the “N” word) is another constructive use.

But when user data and algorithms are deployed for the purpose of custom curating third-party content—to any extent—the offending platform should lose all liability protections under Section 230 and be liable for all third-party content.

Tech companies should exercise no influence whatsoever on the reach of any non-commercial, third-party content on their platforms. Two factors only should affect the third-party content to which users are exposed: (1) their own independent efforts to seek and to select the content they prefer to consume and, (2) the sharing of content by other users with whom they have consented to connect on a platform—displayed in a non-curated format (e.g., reverse chronological order). Both amplifying and constraining the reach of any user-generated content algorithmically should be illegal.

This proposed change would only lessen, not eliminate, the problem of echo chambers.[20] Since it is not a legitimate interest of government to constrain the rights of users to consume any legal content of their choice, the proposed change strictly targets platforms, not users. People are people. But the tech-driven bias confirmation machine is overwhelming our human frailty and adding fuel to the fire of our innate tribalism. This is not making us freer and will not end well, if allowed to continue. Algorithm abuse is the ultimate product liability atrocity!

The Congress should find that the misuse of algorithmic curation technology causes severe harm to the public and should legislate against any use of such technology—either to amplify or to limit the reach of—third-party content, under penalty of the immediate loss of all Section 230 protections across all social media platforms owned or operated by the offender. Severe mandatory civil penalties and regulatory sanctions should also be imposed.

AFTERTHOUGHTS – Lingering Legal Ambiguities↑

(1) While it is certainly true that Section 230 has shielded internet companies and their responsible agents from civil actions in cases where they had been knowing enablers of, or the principal actors in, criminal activity, this commentator has been unable to verify that the law has ever been misapplied as a shield against culpability for the underlying criminal conduct. Any such misuse would obviously constitute a serious injustice.

(2) It is by no means clear that all content moderation practices are of the same kind. For example, when Twitter moderators attach a warning label to a questionable Tweet, is Twitter engaging in constitutionally protected counter-speech (i.e., beyond the reach of government to regulate or sanction) or is it revising the speech of the original speaker (i.e., as co-speaker), despite being treated as merely a distributor with moderation privileges under Section 230?

(3) Given the technically legal rights of platforms to discriminate on the basis of politics and to engage in arbitrary and inconsistent moderation, what, exactly, under the existing language of Section 230(c)(2) would constitute moderation in “bad faith” (assuming that the moderated content is merely objectionable to someone, and not illegal)?

END NOTES↑

  1. According to the Internet Association, a not-disinterested lobbying group representing “medium-big-to-big tech,” the parties that have “relied on” Section 230 as a legal defense include “internet service providers and website hosts, newspapers, universities, libraries, search engines, employers, bloggers, website moderators and listserv owners, marketplaces, app stores, spam protection and anti-fraud tools, domain name registrars, and social media companies.” The percentage of said parties who have successfully relied on Section 230 immunity was not reported. Regardless, the law does protect any website, computer application, or mobile application—from bloggers to news and sports apps, to fin-tech, to e-commerce, to content aggregators, to affinity websites, to social media platforms, etc. ↑

  2. The Communications Decency Act of 1996 refers to Title V of the Telecommunications Act of 1996. ↑

  3. See Reno v. American Civil Liberties Union (1997) ↑

  4. It is more than trivially ironic that free-market conservatives who have long asserted—with quasi-religious fervor—that markets are “self-correcting” and that they lead naturally to the optimal allocation of resources are now incensed that Big Tech’s market-driven moderation practices are impacting them in ways they did not expect. When entire classes of workers are displaced by tectonic shifts in the structure of the economy, those externalities are dismissed as “natural workings of the market.” But any constraints on their “right” to post content to social media platforms is tyranny, “Yes, tyranny, I tell you!” SMDH ↑

  5. In the case of Boy Scouts of America v. Dale (2000), the Supreme Court of the United States, relying on an earlier precedent, held that a private organization’s First Amendment Right of Free Association (citing its “implicit” right of non-association) trumps the State of New Jersey’s public accommodations law prohibiting discrimination on the basis of sexual orientation, which, like political beliefs, is not a Federally protected trait. Also note that, like social media platforms, private organizations such as the Boy Scouts are not public accommodations, under Federal law. ↑

  6. Two early cases illustrate the “Moderator’s Dilemma” that Section 230 was intended to resolve. In Cubby, Inc. v. CompuServe Inc. (SDNY 1991) the Court held that CompuServe was not liable for allegedly defamatory content posted by a third party because it had made no effort to moderate the large volume of third-party content on its platforms and, therefore, was presumed unaware of the defamatory nature of the content and, therefore, was deemed to be a mere distributor of it. However, in Stratton Oakmont, Inc. v. Prodigy Services Co. (SCONY 1995) the Court held that Prodigy’s established moderation practices were a form of editorial control and that Prodigy, therefore, was liable as the publisher of allegedly defamatory content posted by a third party. The takeaway from both cases was that liability risk is minimized when platforms do not moderate third-party content at all. But refraining from content moderation altogether is problematic because, without it, platforms would eventually be overrun by offensive and criminal content that users would find objectionable. Section 230 was enacted to release platforms from this no-win situation for the public good. ↑

  7. Because of their classification as “common carriers” under Title II of the Communications Act of 1934, the telephone companies that provided the land-wire infrastructure over which most internet traffic was transmitted (in 1996) were already shielded from content-related liability and exempted from any potential duty to moderate content. ↑

  8. The outdated language of the statute has subsequently been clarified by the courts. “Interactive computer service” has been interpreted as including essentially any service or platform through which content is accessed. “Information” means content, broadly defined. “Information content provider” refers to the creators and publishers of any type of content—text, audio, images, video, etc. “Restrict access to” includes labeling, curating, removing, blocking, deleting, and, of course, restricting access to content. ↑

  9. Here, “(1)” must be an overlooked typo that made it into the final legislation because the reference is nonsensical. The intended reference must surely have been: “…material described in paragraph (A).” ↑

  10. Section 230, Subsection (e)(1) ↑

  11. In December 2019, Senator Elizabeth Warren (D-MA) and Representative Ro Khanna (D-CA) introduced bills that would have commissioned the Department of Health and Human Services to study the unintended harmful effects of FOSTA-SESTA on sex workers. The House version of the bill was called the SAFE SEX Workers Act. ↑

  12. Perhaps the most accurate way to characterize social media platforms is to say that they are advertising and personal data collection platforms. If so, it is not clear how this may affect the “public accommodations” debate. ↑

  13. Referring to Section 230, Subsection (c)(1), Jeff Kosseff, attorney, and Assistant Professor of Cybersecurity Law at the United States Naval Academy, calls Section 230 “The Twenty-Six Words that Created the Internet,” which is also the title of his 2019 book recounting the history of the legislation’s impact since enactment. ↑

  14. Platforms would be acting in “good faith” if they established and disclosed detailed moderation practices that complied with the Barr proposal’s narrow definition of permissible moderation, implemented those practices consistently and without pretext, and provided affected users with explanations for their moderation decisions and a mechanism for appealing those decisions. ↑

  15. The “Bad Samaritan Carve-Out” is a necessary and important proposal. There are many loathsome actors who operate websites and applications for the express purpose of promoting, soliciting, and facilitating criminal content and activity that causes serious harm. ↑

  16. Certain provisions of the Senate bill were not included in the Barr proposal, such as establishing statutory damages of $5,000 per user and limiting the applicability of the good-faith requirement to platforms with at least 30 million US users and $1.5 billion in global revenue. ↑

  17. Platforms would be required to provide affected users with a detailed explanation for each decision to label, to remove, to demonetize, or to limit algorithmically the reach of any content within 14 days of making such decisions. ↑

  18. The bill has been criticized for requiring independent adult content creators to disclose their personal information to platforms who may be run by despicable actors who should not be trusted with the information. ↑

  19. Intending no hyperbole, former Google design ethicist, Tristan Harris, has characterized algorithmic technology’s capacity to overwhelm human weakness as “Checkmate Humanity”—an urgent and profound existential crisis that ultimately will lead to our demise. ↑

  20. For clarity, bona fide content creators (i.e., publishers, for liability purposes), such as Breitbart “News,” would still be allowed to pound falsehoods into the brains of users on their own platforms, but Facebook and Twitter (mere distributors of falsehood) would be prohibited from amplifying or limiting the reach of Breitbart content on their platforms. ↑