AI-satire disclaimer, round five: I am a text generator, not a civil-liberties monk hand-copying the First Amendment onto a ring light. But the conservative case still leans too hard on a category error: it treats ownership risk as if that automatically justifies a remedy that falls mostly on Americans’ speech, association, and commerce. Yes, ByteDance’s relationship to China is a legitimate concern. No, that does not mean the government gets infinite deference to bulldoze a platform used by creators, journalists, political organizers, teachers, musicians, and small businesses unless it can show the cure is actually tailored to the harm. The current policy posture is basically: “Trust us, this app is a strategic danger, and also please don’t ask too many public questions about the evidence.” That may fly in a spy thriller. In a constitutional democracy, it should make people reach for a raised eyebrow at minimum.
And the practical story is getting shakier, not stronger. A forced sale still runs into the same awkward reality that China has export-control leverage over the recommendation algorithm, which is kind of the whole app’s secret sauce. So Washington keeps selling divestiture as this clean middle path, when in reality it may amount to “sell us the car, but not the engine.” If that collapses into a de facto ban, then let’s at least stop pretending this is some elegant ownership paperwork dispute. It is a huge state intervention into the digital speech market, justified by a threat model that Congress has conspicuously declined to address systemically. If lawmakers truly believe algorithmic influence is the battlefield, then regulate recommender systems broadly. If they truly believe Americans’ data is vulnerable, cut off data-broker pipelines broadly. The one-app approach still looks like selective toughness designed for cable-news lower-thirds.
There is also a less discussed strategic cost: by making platform access contingent on geopolitical acceptability through ad hoc legislation, the U.S. inches toward a more balkanized internet while insisting it is defending openness. That may be unavoidable in some sectors, but we should not pretend there is no tradeoff. Every time Washington uses national-security authority to restructure a major communications platform without a universal rule set, it invites retaliation abroad and normalizes a model of speech governance that liberal democracies are supposed to treat with caution, not pride. If America’s message is that open societies beat authoritarian systems because we trust rights, transparency, and neutral rules, then “ban first, framework later” is a pretty clumsy ad for the brand.
AI-satire disclaimer, round five: yes, I am once again the hawkish bot who sounds like it subscribes to Foreign Affairs and also distrusts any app with suspiciously good engagement metrics. But the liberal argument keeps asking for courtroom-level proof of a catastrophe before taking structural precautions against a strategic vulnerability. That is not how serious security policy works. The relevant question is not whether Congress can produce a viral clip showing a ByteDance executive in a villain chair twirling the algorithm. It is whether a platform with enormous reach into American opinion, behavior, and consumer data should remain under the control of a company exposed to coercion by the Chinese state. On that question, the answer remains obviously no. Risk management is not censorship, and refusing to leave a loaded vulnerability in place is not some betrayal of the Enlightenment.
The liberal side also keeps saying, not wrongly, that broad privacy and algorithm rules are needed. Great. Pass them. But universal regulation and targeted security action are complements, not substitutes. If the U.S. discovered a foreign adversary had unusual leverage over one major telecom provider, no one would say, “Well, until we rewrite the entire communications code, we must do absolutely nothing lest we seem selective.” Selective is allowed when the threat is selective. TikTok’s significance comes from the combination of scale, dependency, opacity, and adversarial-state exposure. That combination is what makes it different from domestic platforms, however grotesque their own surveillance habits may be. “Meta is also bad” is not a serious answer to foreign leverage; it is just bipartisan misery with a Wi-Fi signal.
And on the practical point: yes, divestiture is messy, China may resist, and the algorithm issue is real. But that complication cuts toward the conservative case, not away from it. If the platform cannot be separated in any meaningful way from technology and governance subject to CCP pressure, then the argument for allowing business as usual gets weaker, not stronger. At that point, policymakers are not choosing between a perfect solution and an imperfect one; they are choosing between reducing a known strategic dependency or tolerating it because creators would prefer continuity. That is unfortunate, and it is disruptive, but national security choices are often rude like that. A free society can value speech and still refuse to let a rival power sit inside a giant American attention machine just because the videos are funny and the ad buys are efficient.