The other day the Copia Institute filed an amicus brief at the Fourth Circuit in a case involving West Virginia attempting to punish Apple for not filtering CSAM the way it thinks it should. The issue raised in the appeal was in some respects a narrow one: could Apple cause the case West Virginia brought against it to be heard in federal court, rather than a West Virginia state court, and by using the removal statute at 28 U.S.C. Section 1442(a)(1) in particular. But the implications are much broader than just this particular procedural question. What West Virginia is trying to do, imposing liability on a platform for how it intermediated user expression, interferes with several statutes Congress passed—including Section 230—which sought to shield them from that liability, and for all the sound reasons Congress believed minimizing that potential liability the better policy to pursue.
For one thing, when governments can force platforms to search their systems for evidence of wrongdoing, like through the threat of liability, then the evidence the platforms discover, warrantlessly, will be inadmissible against those users who put it there and thus it will be very hard to successfully prosecute them. The only way to use evidence platforms find is if they find it on their own volition, and that discovery is only volitional when platforms can be legally able to look for it without fearing liability however they do. So Congress passed several statutes to make it safe for platforms to try to police this most odious expression: Section 230, which insulates platforms for liability in most forms wrongfulness they find others used their services for, and the PROTECT Act, which is a separate statute that makes it safe for platforms to police for CSAM in particular.
For West Virginia to now try to impose liability on the platforms because they did not do that policing as it demands means that the results of their search are no longer volitional and instead coerced by the government. As Apple pointed out in its brief, it would make those results inadmissible, and in our brief we showed how a similar principle works in the First Amendment context too, showing how, as NRA v. Vullo explained, the government cannot target someone by targeting someone else they depend on instead. The Constitution will not allow the government to do indirectly what it cannot do directly; if it wants to directly police for CSAM it will need a warrant, as the Fourth Amendment requires, and otherwise, if it wants the help of the platform, as Murthy v. Missouri explained, it will need to make sure platforms are free to offer that help on their own volition.
That Congress wanted these discoveries of CSAM to be prosecutable reflected a significant federal interest, which is undermined by West Virginia meddling trying to impose liability where Congress had specifically imposed immunity. Identifying an “overriding federal interest,” such as ensuring CSAM can actually be successfully prosecuted, is important for being able to remove a case to federal court under that 1442(a)(1) statute. But, as we also argued in our brief, it’s not the only federal interest in play.
With Section 230 Congress also wanted to create an online ecosystem that fostered the most beneficial expression and minimized the most detrimental generally. It did so by purposefully providing platforms with statutory immunity for both its facilitation and moderation of user expression so that they could be safe to do each as best they could. Yet all these state prosecutions of platforms undermine that statutory ecosystem by now imposing liability on the very same functions that Congress sought to protect—functions which, as the Moody v. Paxton case indicated, may themselves be protected by the First Amendment.
The practical freedom Congress tried to afford platforms to do both facilitate and moderate can no longer exist when a platform now has to worry about a state taking issue with how it has done either. Even if the subject of West Virginia’s liability may not represent wrongfulness specifically immunized by Section 230, given the federally criminal nature of its wrongfulness, (1) Congress still immunized platforms for it via the PROTECT Act, and (2) having platforms fear liability for a particular type of wrongfulness makes them fear it for all of it. As we additionally pointed out, West Virginia wanted Apple to look for illegal needles in an enormous haystack of perfectly lawful expression, but if a platform had to fear liability if it didn’t find all the needles as the state wanted it would not be able to facilitate nearly so much lawful hay; it just couldn’t afford the consequences of the government’s disapproval.
Worse, it is not just West Virginia that platforms would have to worry about. State after state has been seeking to impose liability on platforms when they do not intermediate others’ expression exactly how these states wanted, facilitating or moderating either too much or too little for each state’s taste—see, for example, the Florida and Texas efforts behind the Moody v. Paxton case itself. There is always something that some state wants platforms to do more or less of when it comes to intermediating others’ expression, and that something may not even be consistent with what another state wants. Platforms are ending up so pulled into pieces by state regulatory regimes that the statutory protection Congress afforded them is becoming all but useless, because, at the moment, instead of being immune from liability, platforms are ending up targeted by every single local jurisdiction’s.
Which is why Congress had sought to pre-empt this problem through a literal pre-emption provision, telling states (and an infinite number of local jurisdictions) that they were not in the platform regulation business. No state could come up with a regulatory scheme that conflicted with the online ecosystem Congress was trying to foster, one that got the most good and least bad online thanks to immunity that made it safe for every platform to attempt to achieve both. Yet every state trying to impose liability on platforms for what Congress sought to immunize inherently conflicts with it. In fact, it completely obviates the regulatory regime Congress designed because platforms cannot be both free and hindered in how they perform the same intermediating task. All the benefits of immunity are lost once liability is introduced, and all the policy Congress sought to advance with its statutory protection is now subordinated to what individual states would prefer instead.
It reflects a strong federal interest to ensure that the policy Congress sought to vindicate—in clear, unambiguous language—not give way to these state enforcement efforts. At minimum Section 230 is enough to warrant removal into federal court in any of these cases where states are trying to superimpose their policy desires on platforms. But it really needs to start being something that stops them in their tracks altogether.