Ceci n’est pas un Ban?

On 19th January 2025, the ‘Protecting Americans from Foreign Adversary Controlled Applications Act’ (PAFACAA) became operative in the USA in respect of the popular application, ‘TikTok’. Routinely (but somehow deceptively) referred to as ‘TikTok ban’, PAFACAA sparked a heated debate on the relationship between national security and freedom of speech. On 17th January, the Supreme Court upheld the statute’s constitutionality. In a much-publicised move, President Trump subsequently adopted an Executive Order (EO) on 20th January, which purportedly neutralised PAFACAA’s application to TikTok.

PAFACAA and its application to TikTok have been surrounded by considerable controversy and confusion. I will not deal in detail here with the saga (which readers of this blog are already familiar with, thanks to Saxer and Kollenberg). Rather, I aim at highlighting, explaining, and criticising the misalignment between legal form (and its indeterminacy) and political narrative which has characterised the case since its inception.

PAFACAA and TikTok: a primer

In a previous blogpost, I analysed the technicalities of PAFACAA, the background to its adoption, and its systemic significance well beyond TikTok itself. However, for present purposes, I will indeed concisely focus on PAFACAA’s highly-publicised implications for TikTok.

These arise from PAFACAA’s explicit targeting of any ‘website, desktop application, mobile application, or augmented or immersive technology application’ that is operated, ‘directly or indirectly’ by the ByteDance-TikTok group (Section 2(g)(3)(A)). Pursuant to this designation, operators of the US internet infrastructure are prohibited from (Section 2(a)(1)): ‘providing services to distribute, maintain or update [TikTok] by means of a marketplace’, and ‘providing internet hosting services to enable the distribution, maintenance, or updating of [TikTok]’.

As is by now well-known, this could have been avoided if a ‘qualified divestiture’ had been made before 19th January 2025 (Section 2(c)(1)(A), in conjunction with Section 2(a)(1)(A); see Section 2(g)(6) for the definition of ‘qualified divestiture’), and PAFACAA will likewise cease to apply if a qualified divestiture takes place anytime in the future (Section 2(c)(1)(B)). In other words: either TikTok severs its ties with China, or US operators must stop providing services which directly or indirectly allow for TikTok to be ‘distributed, maintained, or updated’ in the US. Failing a qualified divestiture, entities providing the prohibited services face draconian financial penalties to be enforced by the US Attorney General (Section 2(d)).

An indirect TikTok ban?

Ever since its adoption, PAFACAA has been narrated by the media as a ‘TikTok ban’ (see e.g. prominent outlets in the US, Germany, France, Italy, and the UK). Yet, as more recently acknowledged by the press itself (see e.g. here, here, here, and here), PAFACAA’s actual content is much more indeterminate than the ‘ban’ discourse might suggest.

Following Lessig’s influential theorisation (see 129-136), a ‘ban’ on TikTok could be direct or indirect. A ‘direct ban’ would entail, for instance, placing on TikTok itself a sanctioned obligation not to operate the application. This route was taken by a controversial 2023 Montana statute, which was quickly found, inter alia, to impermissibly restrict the freedom of speech under the First Amendment. Mindful of this precedent, PAFACAA ostensibly refrained from adopting this radical approach. However, TikTok could also be subject to an ‘indirect ban’: law could regulate the broader technological environment so as to effectively prevent TikTok’s operation. TikTok was also subject to this strategy in 2020, when President Trump adopted Executive Order (EO) 13942. As fleshed out in its implementing act, EO 13942 prohibited US companies from providing a whole range of services to TikTok, whenever those services would assist in TikTok’s ‘distribution’, ‘maintenance’, ‘updating’, ‘functioning’, or ‘optimization’. This effectively meant that TikTok would have had to rely on its own technical capacity to deliver content to each of its 170 million US active users. As TikTok did not dispose of such titanic capacity, EO 13942 did effectively amount to an indirect ban on TikTok. EO 13942 was found to prima facie exceed the powers delegated to the President by the underlying statutory authority. However, an equivalent regime parallelly targeting ‘WeChat’ was found to prima facie breach the First Amendment, precisely because of the indirect ban it brought about.

A comparison with EO 13942 evidences PAFACAA’s softer regulatory touch. Both acts prohibit US companies from servicing TikTok when the services are functionally linked to a specified range of activities. In PAFACAA’s case, these are (only) the ‘distribution, maintenance, and updating’ of TikTok – activities which EO 13942 prohibited, too. At a minimum, this entails the following. By preventing TikTok’s ‘distribution’, PAFACAA prohibits service providers from making TikTok available for download, thus preventing new users from joining the platform.

However, as mentioned, TikTok already has 170 million active users in the US, in respect of whom this prohibition is in principle without implications. It is here that PAFACAA’s two other prongs enter the scene. By targeting TikTok’s ‘maintenance’ and ‘updating’, PAFACAA also prohibits service providers from allowing users who already installed TikTok on their devices to download bug fixes and updates. Admittedly, this prong of PAFACAA may, in the medium-to-long-term, amount to an indirect ban. In fact, this will probably induce a technological obsolescence of TikTok. That is, TikTok will end up not being concretely usable, as its operability on those devices (which will keep on routinely being updated) slowly diminishes and cybersecurity loopholes in the app cannot be repaired any longer.

However, the extent to which PAFACAA brings about an indirect ban in the short term is at least ambiguous. In fact, unlike EO 13942, PAFACAA does not prevent US companies from also contributing to TikTok’s ‘functioning’ and ‘optimization’. This is a crucial difference. Prohibiting these activities undisputedly made it illegal, under EO 13942, for US operators to allow TikTok to distribute or store information through their data centers, content delivery networks, or transit providers. This surely entailed the indirect ban chastised in the WeChat case as a breach of the First Amendment. The lack of a comparably explicit provision to this effect in PAFACAA makes the latter’s immediate implications for TikTok less self-evident. To be sure, Congress’s deliberations were taken under the assumption that PAFACAA did indeed entail an immediate indirect ban. However, the extent to which this is reflected in the final statutory language is unclear. This should be problematised rather than glossed over. It could perhaps be maintained, placing a premium on Congressional intent, that the prohibition for internet hosting service providers to ‘enable the distribution’ of TikTok entails a broader obligation not to deliver TikTok’s content through one’s infrastructure (as Rozenshtein’s sophisticated analysis seemingly implies). However, as evidenced by the comparison with EO 13942, this conclusion is not compelled by PAFACAA’s language. Therefore, PAFACAA may as well be taken to only have the mid-term implications highlighted above (as spotted early-on by Ball: see 2:36-3:23).

PAFACAA’s ambiguity could very well be a deliberate attempt to achieve the purpose of indirectly banning TikTok without engaging upfront with the First Amendment challenges entailed by EO 13942’s more explicit provisions. In fact, uncertainty over the scope of application of US sanctions programmes, from which PAFACAA draws inspiration, systematically leads to the well-studied phenomenon of ‘overcompliance’. Operators of the infrastructure being weaponised (in sactions programmes, essentially, the banking infrastructure) routinely carry out way fewer transactions than they arguably could, fearing to incur hefty penalties under regimes whose scope is unclear by design. Against this background, PAFACAA could have deliberately been couched in ambiguous terms to induce the US internet infrastructure into overcompliance (as suggested by the case of Oracle, which however eventually kept on servicing TikTok). This issue should also be problematised on its own right. However, these subtleties have largely not featured in the discourse surrounding PAFACAA. Rather, this by and large stuck to the immediate ban narrative.

The turn to (legal and political) symbolism

Perhaps surprisingly, even the legal battle over PAFACAA mostly embraced the ban narrative. In their petition for review of PAFACAA’s constitutionality, TikTok and its parent company, ByteDance, generically alleged that PAFACAA prohibited TikTok from being ‘serviced’ and that this prohibition merely ‘included’ services assisting ‘the distribution, maintenance, or updating [of TikTok] through an online marketplace’ (para. 31). The submission thus suggested (but also did not state explicitly) that the range of services encompassed by PAFACAA’s prohibitions was in fact broader.

However, in their subsequent Brief, petitioners appeared even more hesitant. On the one hand, they again assertively maintained, in line with PAFACAA’s clear baseline, that ‘operators of an “app store” may not allow users to download or update [TikTok] in the United States’ (Section B.1). In outlining PAFACAA’s further implications, however, petitioners more speculatively maintained that ‘data centers would almost certainly conclude that they can no longer store [TikTok’s] code, content, or data’ (ibid., emphasis added). That this conclusion was far from unambiguous to petitioners themselves is also evidenced by the embarrassed hesitancy with which TikTok’s counsel orally reiterated this line of argument before the Supreme Court (see 50:46-52:22).

Such ambivalence probably was an understandable strategy to subtly instill the idea that PAFACAA indeed entailed an indirect ban, exercising however a rhetorically wise caution. It is somehow more puzzling that counsel for the US government itself stuck to this framing. In fact, the Respondent Brief (para. 3) argued that:

‘[t]here is no dispute in this litigation that the deprivation of [the services enabling TikTok to be distributed, maintained, or updated] would practically preclude [TikTok] from continuing to be widely offered to American users’.

While ambiguity could still exist as to the precise meaning of ‘widely offering TikTok to American users’, the Brief’s overall tone is indeed aligned with the ‘ban’ narrative. This unreceptiveness to nuance seems counterproductive: as shown by the WeChat litigation, framing PAFACAA as an indirect ban arguably makes it problematic under the First Amendment.

This lack of strategic thinking probably has to do with the symbolical meaning soon taken on by the whole case. In the public debate, the contest around PAFACAA soon became a proxy for the long-standing confrontation between freedom of speech and national security. It is thus unsurprising that this symbolical framing spilled-over into court proceedings. These soon started to revolve around questions of principle, obfuscating discussions over legal nuance. Amici briefs siding with petitioners were filed, inter alia, by free speech activists and right-libertarian think-tanks, whereas members of the national security community and Chinese dissidents supported PAFACAA’s constitutionality. As the discourse became ever more polarised, attention to PAFACAA’s concrete working faded into the background, and legal as well as political arguments progressively took for granted that PAFACAA did indeed entail an immediate indirect ban. In fact, this narrative even made it in the Supreme Court’s judgement, which simply assumed that PAFACAA was an ‘effective ban’ on TikTok (Section II.A; although see Justice Gursoch’s Concurring Opinion, acknowledging that ‘what might happen next to TikTok remains unclear’).

PAFACAA’s progressive radicalisation in the public discourse was compounded by TikTok’s spectacular decision to voluntarily shut down the day before PAFACAA’s entry into force. Users trying to access the app were shown a message stating that ‘[a] law banning TikTok has been enacted in the U.S. Unfortunately, that means you can’t use TikTok for now’. In so doing, TikTok effectively suggested that PAFACAA even brought about a direct ban. Far removed from legal reality, the move was clearly an attempt to exercise political pressure on the US government. This was made apparent by the message’s specification that TikTok was ‘fortunate that President Trump has indicated that he will work […] on a solution to reinstate TikTok once he takes office’.

The fiction went on with Trump’s EO, presented to the public as a (temporary) neutralisation of PAFACAA – a congressional statute, clearly not amenable to such executive impairment. In fact, the EO much more soberly directed the Advocate General not to enforce PAFACAA’s prohibitions for 75 days: a ‘solution’ which does not substantively prevent service providers from incurring liability if they keep on allowing TikTok to be ‘distributed, maintained, or updated’ (see Rozenshtein). However, TikTok perpetuated its misrepresentation of reality, suggesting not only that PAFACAA had put in place a ‘ban’, but that Trump was to be credited for lifting it. In fact, after the EO’s adoption, TikTok went back online (which it was never prevented from doing, either de iure or de facto), sending users a cheerful in-app message that ‘[a]s a result of President Trump’s efforts, TikTok is back in the U.S.’. At the same time, PAFACAA’s ‘real’ addressees are well aware of the EO’s limits: at the time of writing, TikTok, other applications run by the ByteDance group, as well as updates/fixes thereto are indeed not available for download on US app stores. As hereby argued, this is PAFACAA’s only completely unambiguous implication – and one which, for all the narrative on Trump’s rescue of TikTok, is fully displaying its effects.

The responsibility of legal scholarship

The TikTok-PAFACAA saga illustrates the complex relationship between law and public discourse in contemporary debates. A vaguely formulated statute became a symbolical proxy for principled confrontation over the underlying values. In the process, PAFACAA’s probably deliberate ambiguity got lost in translation. This radicalised narrative allowed for an unlikely alliance between TikTok and Trump, formerly TikTok’s staunchest opponent. This awkward covenant was cemented in an EO of dubious legal soundness, deceptively presented as the last bastion of free speech in the US.

Of course, Trump has strong political incentives to U-turn on TikTok. In turn, TikTok has strong economic incentives to cherish its relationship with an administration in which the historical collusion between Big Tech and the US political establishment has reached unprecedented levels. Against this background, legal scholars can play a role aligned with what thinkers of varied ideological inclination have more broadly theorised for intellectuals in contemporary societies. In fact, I submit that, regardless of one’s stance on ‘scholactivism’, legal scholarship has at least a minimal responsibility to (try to) contribute to societal debates. By attempting to elucidate the state of the law and its (deliberate) ambiguities, scholars should expose, and thereby open to scrutiny, the strategic motives which lead to an oversimplified public discourse.