Former Massachusetts Attorney General Scott Harshbarger and former San Francisco Chief Assistant City Attorney Dennis Aftergut recently made a fascinating proposal on Just Security, arguing that state attorneys general (AGs) should lead litigation against Big Tech, as they did with Big Tobacco 23 years ago. While the proposal offers an attractively creative method of regulating Big Tech and holding it accountable, a few cautions are warranted – namely that, in today’s hyper-polarized political environment, empowering elected, partisan AGs to dictate the way in which information on platforms circulates likely comes with more costs and fewer benefits than the authors assume.
Big Tech and Big Tobacco May Not Be as Similar as They Seem
The Harshbarger and Aftergut proposal seeks to solve harms ranging from social media’s disastrous effects on youth mental health to disinformation by harnessing national coordination between all 50 AGs, akin to the litigation strategy that resulted in the Tobacco Master Settlement Agreement (TMSA) of 1998. Scholars have called the TMSA, which required the largest tobacco companies to pay over $200 billion and included every state, the functional equivalent of federal law. This nationally coordinated AG model, the authors argue, should now be transplanted to another Big T – Big Tech. They further argue that courts ought to recognize an “AGs enforcing state law” exception to the (in)famous Section 230 of the U.S. Communications Decency Act (47 U.S.C. § 230, or “Section 230”), which currently immunizes platforms like Facebook from harms that result from both content Facebook lets circulate and its moderation decisions.
Harshbarger and Aftergut’s proposal is timely. The Ohio AG recently became first to file a suit against Facebook modeled in part on the tobacco claims. Shortly thereafter, a coalition of AGs from 10 states launched a joint investigation into Instagram’s impact on teens. Harshbarger and Aftergut are very likely correct that AG suits against Big Tech will increase, prompting inevitable comparisons to tobacco suits.
As the authors point out, too, in consumer protection suits against both the tobacco companies then and tech companies now, private litigation utterly fails to generate compensation or corporate accountability. And in both cases, Congress failed to update the existing statute or pass new ones to address novel but obvious public policy problems. Those two factors make AG suits, as an option of last resort, enticing.
But private claims failed in the two contexts for very different reasons. The tobacco defendants attacked the victims as irresponsible addicts – and, employing a “blame the smoker” strategy – had a near-perfect record until the AGs got involved. As the architect of the AGs’ litigation strategy, former Mississippi AG Mike Moore, put it, the state had “never smoked a cigarette,” and so couldn’t be blamed.
Big Tech is different. Private lawsuits against platforms have failed (or are highly likely to fail) due to a federal statute known as Section 230. Section 230 essentially immunizes platforms from civil lawsuits based on real-world injuries that their user-generated content causes by treating the platforms as publishers, not speakers, even in the context of algorithmically “boosted” messages, and immunizes any “good faith” content moderation choices that the platforms do undertake, on the same publisher theory.
As the authors rightly acknowledge, for the AGs to sue the platforms regarding any of the regulatory issues they identify, they would need to get a court to hold that Section 230 didn’t apply to AGs ever when they litigated state law claims regarding any asserted “informational” harm. And they propose a novel “loophole” to do just that:
Liability is not barred for actions under state law that are “consistent with” Section 230. A strong argument is available that a[n AG] lawsuit against Facebook based on state consumer laws that prohibit deceiving the public is “consistent with” §230’s immunity for a “publisher” of content.
Although novel, such a reading is far from implausible. While historically the Supreme Court has read Section 230 broadly, that appears very likely to change as the balance of the Court shifts in favor of conservative justices. In 2020 alone, Justice Thomas publicly inveighed on Section 230’s breadth twice, explicitly castigating how it legitimizes Twitter’s power to “cut off speech” as against core First Amendment values. That position is echoed by conservative jurists on lower courts, most recently DC Circuit Judge Laurence Silberman, whose inflammatory dissent charged Section 230 with permitting powerful Silicon Valley firms to engage in extreme “repression of political speech” which is “fundamentally un-American.” Prominent conservative jurists are actively searching for a mechanism to cabin Section 230’s reach – and an AG exception based in principles of states’ rights and state sovereignty may be an appealing means solution. The Court has consistently afforded AG litigation with what it has called a “special solitude” based on strong deference to sovereign enforcement of state laws, permitting AG suits to bypass other limitations for antitrust enforcement, tobacco litigation, and, most recently, opioid litigation.
Still, the novelty of this argument will likely make it an uphill battle. And beyond the question of whether courts would recognize the “AG exception” to Section 230 is the equally important question of whether they should recognize it.
The objective of the Aftergut-Harshbarger proposal – utilizing tobacco-style suits to regulate Big Tech’s enabling of viral spread of misinformation, which endangers everything from public health to democracy itself – is beyond reproach. But it overlooks three related problems with translating the AG model from Big Tobacco to Big Tech. Taken together, these problems suggest that the costs of recognizing an AG exception to Section 230 outweigh the benefits.
What Benefits Would AG Litigation Actually Deliver?
Overall, the long-term benefits of the TMSA have been disappointing, which bodes poorly for the real-world impact of consumer litigation that takes tobacco as its model.
First, Aftergut and Harshbarger argue that transposing the TMSA model would produce financial compensation for Big Tech’s victims. But while AG Harshbarger’s state of Massachusetts used settlement money to compensate victims and fund public health initiatives, a large majority of other states did not. According to a recent article by leading public health litigation experts Robert Rabin and Nora Freeman Engstrom, the TMSA failed to compensate victims structurally: the settlement contained no requirements on how states would spend their recoveries, so the money was deposited in state general treasuries and used for whatever purpose the legislature saw fit. The result, according to Engstrom and Rabin, was that states “cannibalize[d]” the TMSA money for uses entirely unrelated to public health, let alone tobacco.
Criticism of the TMSA has come from all corners. Senator John McCain took the AGs to task for how they used the TMSA in 2003, as did numerous media reports in the early aughts. As of 2016, the America Lung Association, the main consumer advocacy group tracking use of the TMSA funds, gave 41 states an “F” because, even still, virtually none of their recovery had gone toward its intended purpose. The GAO and NIH’s assessments are equally dire, while TMSA architect Moore called the way funds were allocated one of the “biggest disappointments” of his career.
Second, and perhaps more importantly, the authors argue that the TMSA model provides a long-term and effective mechanism to regulate Big Tech. But the TMSA’s regulatory and deterrent effects on the tobacco industry were largely ineffective, and in some ways counter- productive. That settlement turned the 50 states into the biggest shareholders of the five largest tobacco companies, which some argue functionally immunized the tobacco industry from antitrust claims while locking in the Big 5’s market dominance. The Ninth Circuit expressly rejected antitrust suits despite evidence of a price-fixing conspiracy, holding they were preempted by the TMSA. The antitrust issue is more problematic in the Big Tech context where antitrust has emerged as the key vehicle of corporate accountability.
The upshot is that evidence from the vast majority states shows a TMSA approach would result neither in compensation for victims nor effective regulation of Big Tech.
AG Litigation Risks Further Polarizing Big Tech Regulation
The nationwide coordination that made Big Tobacco litigation successful would also be necessary for Big Tech litigation to work, as Aftergut and Harshbarger acknowledge. But, given the extreme levels of political polarization regarding Section 230 and Big Tech, it is more likely that AGs today would fracture along party lines and fail to coordinate nationally, which would balkanize internet regulation and feed bitter cycles of polarization.
As Margaret Lemos points out, one of the most important considerations about AG suits is how any such litigation would fit into the politics of the moment. This is because the AGs are not normal private lawyers or civil servants at the Department of Justice or Federal Trade Commission. Rather, they are majoritarian political actors – 43 of 50 are directly elected – who use their position as a springboard for higher office often enough to generate the tongue-in-cheek aphorism that “AG stands for almost governor.” As such, one crucial driver of their litigating, settling, and fund distribution choices isn’t altruism – it’s political self-interest. After all, the main reason most elected officials do just about anything is to win more elections.
Big Tobacco was a (much) less polarized policy issue litigated at a (much) less polarized time. To accurately assess what AG litigation would look like if applied to Big Tech, the first step therefore ought to be a close look at the political incentives AGs face in the current political environment.
Politics is determinative. Lemos and Ernest Young in a subsequent article draw a useful distinction between “horizontal” and “vertical” AG litigation, arguing that only the latter is ripe for a national AG strategy in an age of extreme polarization.
The paradigm cases of vertical litigation would be those like tobacco and opioids, the most recent attempt to transplant the TMSA model, in which broad agreement among both parties over the policy issue makes national coordination likely. With polarization becoming “existential” in the United States, tobacco and opioids are unique in that supermajorities of Americans, irrespective of ideology or party, agreed litigation should punish Big Tobacco and Big Pharma. These vertical suits have positive value – they represent the states acting collectively to take back some control from the federal government and can even help depolarize the country.
But during both the Trump and Obama presidencies, AGs found a new way to use litigation to advance their political ambitions – to litigate as partisans around a polarized issue. Lemos terms this “horizontal” AG litigation, resulting in conflicts between the states, bifurcated by party alignment. Under Obama, Republican AGs became exponentially more litigious, using courts to resolve deeply polarized political questions, including climate rules, the Affordable Care Act (which they succeeded in partially striking down), and school bathrooms. Increasingly, these AGs started to see horizontal litigation as part of their job descriptions. Democratic AGs further accelerated the rate of such partisan suits under Trump, suing the administration 138 times, nearly double the rate under Bush and Obama.
The degree of polarization between Democratic and Republican AGs is becoming both more pronounced and more dangerous during the Biden presidency. Look no further than Texas’s suit requesting the Supreme Court enjoin counting Pennsylvania’s Electoral Votes based on false conspiracy theories about widespread voter fraud. Though unsuccessful, the Texas litigation illustrates how political incentives generated by polarization drive the AGs to litigate: Texas AG Ken Paxton reaped enormous political benefits like fawning conservative media coverage and effusive praise by Trump himself. Eventually nearly all Republican AGs joined the litigation, while nearly all Democrats filed briefs in opposition. Merely by filing the case and bringing the horizontal conflict in public discourse, Texas v. Pennsylvania damaged the legitimacy of American elections, feeding into the destabilizing violence after the election. The trend is persisting with AGs litigating challenges to OSHA testing or vaccination rules for businesses.
Does AG litigation against Big Tech more parallel the bifurcated horizontal litigation of Texas, or the united, vertical litigation that generated the TMSA? Almost certainly, the answer is horizontal – and highly polarized.
After Trump’s ban from major tech platforms, policy debates surrounding the regulation of Big Tech, censorship, misinformation, and Section 230 have become some of the most polarizing issues in Washington. Republicans broadly believe Big Tech should be punished for its liberal bias and censorship of conservatives. Multiple GOP-controlled state legislatures have introduced so-called “anti-censorship” laws, attaching liability to what the platforms did in the aftermath of the January 6th attack on the Capitol. Even before his own ban, Trump threatened to veto a critically important defense spending bill because, as Trump Tweeted, so-called “Big T” used Section 230 to defend removing dangerous election conspiracies. Because 90% of Republicans believe the platforms unfairly censor conservatives, it’s no surprise that the crux of the GOP’s Big Tech policy concerns stripping the companies from the power to removing disinformation. This is no minor issue for Republicans, who increasingly view Section 230’s permissive content moderation regime as amounting to Orwellian censorship and outright liberal authoritarianism.
Extreme polarization follows because, while Democrats agree that Big Tech is a problem, they think so for the opposite reason. Especially after Jan. 6, Democrats have aggressively demanded altering the Section 230 regime to require greater moderation of far-right content. Democrats think the platforms are far too permissive of violent, extremist, inciteful content and hate-speech to circulate and that, if anything, banning Trump’s accounts was too little and too late.
AG Litigation Would Concentrate Power in the Hands of Hyper-Partisan Litigants
Entrusting regulation of Big Tech to hyper-politicized actors – increasingly only accountable their parties’ fringe bases in primary elections – is dangerous. It would empower some of the most partisan litigants to regulate the flows of information throughout the country and world.
This concern most obviously implicates Aftergut and Harshbarger’s proposal that AG suits should be the means to tackle regulation of misinformation. But it also applies to other examples, such as suing Big Tech for boosting content allegedly harmful to minors’ mental health. To be successful, any such litigation would require a court to recognize the public interest exception to Section 230 for AG suits about informational harms. A holding opening the door to any such suits would almost certainly open the door to other AG actions purportedly in the public interest. Once the AGs can bypass Section 230 for some informational harms, they can do so for all alleged informational harms. Given the incentives, it would be a matter of time before politically motivated AGs used the exception in damaging ways.
A Section 230 exception would permit the AGs to sue to enforce, for example, one of the state laws banning content moderation of conservatives. What if an AG – like those shameless enough to pursue flagrantly false election fraud claims – sued under such an anti-censorship state law and got an injunction requiring Facebook to rescind its ban against Trump? The AG’s injunction would result in a legal right for Trump to spew conspiracy theories, call for further violence against Democrats and so-called “RINOS,” and otherwise enflame his base with “stand back and stand by” rhetoric. (Note that this would actually be a relatively benign use of the AG exception – an AG could also seek to restore Neo-Nazis removed after Charlottesville, participants in the January 6 attack and militias plotting the violent overthrow of the federal government, under the same theory.)
In that scenario, the obvious (and only) response from a Democratic AG – representing a constituency with equally strong views on content moderation – would be to also obtain an injunction. But this one would require Facebook to ban Trump to prevent circulation of misinformation under their state’s laws. The result, legally and pragmatically, would be uncertainty verging on anarchy. Should the court somehow split the internet in two, reflecting the kind of separation that exists between Chinese and American cyberspace? Such an outcome would guarantee the cacophonous fragmentation that, as Richard Pildes demonstrated, collapses the very possibility of coordination as such in national governance. After all, how could a court possibly choose which injunction to uphold and which to strike down? There would be no neutral way to make the choice because, under the doctrine of equal sovereignty, each state’s right to enforce its own laws is equivalent to all other states’. Any choice would spur more polarization and greater political fragmentation while delegitimizing the judiciary as nakedly partisan.
Because Big Tech is not Big Tobacco and 2021 is not 1998, all proposals for importing the AG model today must grapple with the political complexities of the present.