RECKLESS ASSOCIATIONS.

• Author: Bambauer, Jane R.

TABLE OF CONTENTS I. INTRODUCTION 488 II. THE AMBIGUOUS RIGHT TO FREELY ASSOCIATE 494 A. Group Theory 494 B. Strict in Theory, Looser in Fact 498 C. Toward a Theory of Secondary Liability 501 D. Proceeding with Caution 503 III. RADICAL ASSOCIATIONS 506 A. Associations Before and After the Internet 506 B. The Mixed Effects of Radical Associations on Beliefs 511 C. From Radical Associations to Physical Harm 517 D. Strained Legal Theories 520 IV. RADICAL ASSOCIATIONS AND TORT FIRST PRINCIPLES 525 A. The Compensation and Harm Principles 525 B. Filling Pockets of Underdeterrence 529 C. New Opportunities for Ex Post Enforcement 532 V. SECONDARY LIABILITY FOR RECKLESS ASSOCIATION 533 A. The Elements 533 B. Sufficient Causation 534 C. Sufficient Mental State 540 VI. OBJECTIONS 544 A. Chilling Effects 544 B. A Deluge of Disclaimers 547 C. Biased Enforcement 547 VII. CONCLUSION 549 I. INTRODUCTION

Social media platforms enable like-minded individuals to find each other and communicate freely about shared interests. This networking is the reason social media exists, and it works well for many purposes--entertainment, shopping, and general social interaction. (1) But it has come at a cost. Individuals who share preexisting political beliefs can become radicalized and emboldened. (2) Predictably, some of them will commit extreme and harmful acts. There is no better symbol of this problem than the crowd who stormed the U.S. Capitol on January 6th, 2021, after years of social network communications had left them angry or deluded about the integrity of U.S. institutions and elections. (3)

Legal scholars and policymakers are exploring new types of law that might mitigate these risks without impeding the beneficial qualities of social media, but so far the interest in new forms of regulation has focused primarily on platforms. (4) This Article takes a different tack. We discuss how legal liability can be responsibly extended from radicalized individuals who physically assault officers, counterprotesters, and bystanders, to the central nodes of their social networks. These individuals, such as former President Donald Trump and the less well-known influencers of fringe groups like the Proud Boys, Antifa, the Oath Keepers, and QAnon, are the actors within a complex communications ecosystem with the greatest moral responsibility and practical ability to deter violence. (5) They are the most trusted and influential nodes in radicalized networks and are, therefore, the best subjects of legal deterrence. (6)

While leaders of radicalized networks have been the subject of civil lawsuits and criminal prosecutions, (7) there has been surprisingly little attention paid to the deep legal and policy considerations related to holding individuals responsible for crimes carried out by other members of their groups. To be sure, mainstream media has accurately portrayed President Trump, Alex Jones, "Q," and others as recklessly indifferent to the cumulative impact of their online speech, (8) but those observations have not carried over into a policy conversation about expanding or revising the law. Theories of legal responsibility being tried in court--namely conspiracy and incitement--are likely to be difficult and will often fail when applied to group leaders who were not giving explicit orders in real time or themselves committing crimes. (9)

As a result, leaders of radicalized groups can skirt liability by restricting their words and deeds so they can deny sharing a specific criminal end goal. The central nodes in radicalized networks can avoid doing the "dirty work"--attacking Capitol Police officers, (10) or intentionally driving a car into a group of counterprotesters (11)--while fostering communication environments that contributed to those attacks. Worse, leaders are in the best position to shield themselves in this way. The de facto leaders of radicalized groups are able to avoid legal risks even though the real-world dangers that their persistent media practices created may have been obvious to them. They observe members of their unofficial associations sink slowly into irreversible paranoia. While true believers go to jail for actions taken in the real world, the self-serving leaders of their movements can guard themselves with plausible deniability. (12)

Thus, there is good reason to believe that the current tort system underdeters behavior that poses unjustified risks to society. We offer a remedy--a new tort--that imposes secondary liability on individuals who assume a position of influence within a radicalized network that recklessly causes network members to physically harm others.

To illustrate: Alex Jones could be held civilly liable for the physical harm caused by the shooting at Comet Ping Pong pizzeria (13) if the victim, through the discovery of Twitter, Facebook, and other social networking data, could show that Jones was among the most influential nodes in the shooter's network that persistently trafficked in "Pizzagate" pedophile conspiracy theories. (14) Jones could exercise a defense based on lack of sufficient mental state (reckless indifference) if he could show even a modest attempt to correct the record or dampen the potential hostility before the incident--a defense we believe he would not actually be able to muster. (15)

This form of liability we call Reckless Association is simultaneously modern and traditional. It is modern because it uses forms of evidence like communications metadata, network analyses, and machine learning that are a product of the new information age. These forms of evidence will be critical for a plaintiff who must prove that a defendant caused an attack to occur and was sufficiently aware of the risk. But Reckless Association is also traditional because it is grounded in political theories that balance duties and liberties across members of society and go no further than necessary to reduce unjustified risks. (16)

Scholars and policymakers are aware of the adverse effects of radicalized online networks. (17) But policy proposals are often directed at social media companies, rather than network leaders, on the theory that platforms amplify or fail to remove radical content (from a progressive perspective) or exacerbate polarization by engaging in biased content moderation (from a conservative perspective). (18) If policymakers are ready to blame Facebook and Twitter for this fiasco, the de facto leaders of radicalized, violent groups are more blameworthy by having better insight into the customs, beliefs, and motivations of their communities and actively encouraging a grievance mindset. And yet, there has not been a serious effort to apply liability to these leaders unless their posts and offline conduct satisfy the requirements for incitement, conspiracy, defamation, or sedition.

We suspect this omission results from tacit assumptions and misconceptions about the First Amendment. The implicit logic of contemporary debate is that courts cannot reach the central nodes of a radicalized network without causing a chilling effect that would inhibit speech and free association. While this is true in one sense--liability may very well cause individuals to avoid becoming authority figures in groups that aggressively traffic in conspiracy theories and radicalizing rhetoric--if liability is appropriately constrained, it should meet the requirements of constitutional scrutiny for the same reasons that narrow versions of defamation and incitement law do. (19) Moreover, to the extent the First Amendment constrains liability based on Reckless Association, it would likewise constrain laws that would force major platforms to purge content and users. (20) Policy remedies that target social media platforms are likely more offensive to First Amendment values because if platforms face liability risks, they are bound to overcensor. (21) It's difficult for outsiders to predict which groups pose a risk since no one platform can see coordination and communications taking place in other online and offline fora. Platforms will not be highly motivated to keep censorship to a minimum if there is a significant risk of liability for only one form of error (wrongly permitting misinformation to be communicated).

That said, liability for Reckless Association must be designed carefully to fit the spirit and doctrine of First Amendment protections. After all, we propose secondary liability based entirely on a defendant's role in an expressive association. Nevertheless, liability can be well-tailored to a serious risk to society by requiring physical injury and proof that the defendant was a very active and persistent central node in a radicalized communications network. A bounded tort of this sort should be able to withstand scrutiny. It would be narrow enough to avoid unnecessary chilling effects but would also be more flexible and fitting to the risks of our modern hyper-networked communications ecosystem than existing forms of liability.

Given the uncertainty about how the First Amendment will constrain theories of liability premised primarily on patterns of association, we begin in Part II by elucidating the individual right to freely associate on social media. Part III explains how the radical freedom to associate, made possible by social media, has brought tremendous benefits as well as some risks to society. The risks are predictable given the theory and available evidence about how people form beliefs and act under the influence of highly selective, frictionless associations, but they nevertheless fall into a liability vacuum. Part IV explains why secondary tort liability, imposed on central nodes of a radicalized network, is the most incisive response to the problem, and it shows that such liability is consistent with the goals and purposes of tort law. Part V serves the main course: it lays out the elements for a new tort of Reckless Association and explains how two of the most important...