The European Commission has formally charged TikTok with failing to protect users’ mental health under the Digital Services Act — specifically targeting the platform’s “For You” recommendation algorithm, which regulators say is designed to maximise addictive engagement at the expense of user wellbeing, particularly among minors. The charges, if upheld by the Commission’s enforcement process, could result in fines of up to 6% of TikTok’s global annual turnover, and the EU has indicated it may require structural changes to the algorithm itself rather than merely additional user controls. This follows a previous probe into TikTok’s data handling practices and represents the most aggressive regulatory action under the DSA since it came into full force. TikTok, owned by ByteDance, rejected the characterisation of its algorithm as inherently harmful and said it would contest the findings.
The received wisdom
The mainstream progressive case for this intervention is both intuitive and backed by a growing body of research. Adolescent mental health has deteriorated significantly in the smartphone era, with correlations between heavy social media use and rates of depression, anxiety, and self-harm — particularly among teenage girls — documented in the work of Jonathan Haidt and others. TikTok’s algorithm is, by design, more powerful and more effective at capturing attention than earlier social platforms: it requires no social graph, learns user preferences at extraordinary speed, and serves an essentially infinite content feed tuned to psychological vulnerabilities. The EU’s argument is that this constitutes a systemic risk that individual users cannot meaningfully consent to — that a teenager cannot realistically opt out of an algorithm optimised to make opting out feel impossible. Regulators, on this view, are filling a market failure. Brussels is doing what Washington has been unable to do because of First Amendment constraints and Congressional deadlock. That case deserves genuine respect.
A different read
The problem is not with the diagnosis but with who is being handed the prescription pad. The EU’s DSA framework is genuinely novel: it attempts to regulate not just illegal content but “systemic risk,” which is defined in ways that give the Commission wide discretion to determine what kinds of online experience are acceptably harmful. That discretion matters enormously, because the same framework that today targets TikTok’s addictive scroll could tomorrow be used — and is already being discussed — in reference to news aggregators that surface “disinformation,” political speech that is deemed to create “polarisation,” or content that contradicts official public health guidance. The slippage from “children’s mental health” to “democratic discourse” is not a hypothetical: the Commission has already raised questions about algorithm-driven political content in the same DSA proceedings.
The right-of-centre tradition has always been wary of paternalistic regulation that begins with a compelling welfare case and ends with the state determining the acceptable shape of public discourse. That wariness does not require defending TikTok’s business practices, which are genuinely predatory in design. It requires asking: who decides what an algorithm “should” look like? If the answer is the European Commission — an unelected executive body accountable primarily to member state governments rather than to users — then the cure may be structurally worse than the disease. The history of state-directed media regulation does not inspire confidence. The British system of press regulation produced the Leveson inquiry and decades of politicians attempting to co-opt the press complaints process. The German approach to platform liability (the NetzDG law) produced mass over-removal of legal content as platforms chose deletion over legal risk.
There is also the specifically geopolitical dimension that the mainstream framing tends to soft-pedal. TikTok is owned by ByteDance, a Chinese company subject to Beijing’s national security laws, which require cooperation with intelligence services on request. The EU is, on its current course, seeking to regulate TikTok’s algorithm for mental health reasons while simultaneously entertaining ByteDance’s lobbying about the form that regulation should take. The two concerns — algorithmic harm to children, and potential CCP data access to European users — are related but distinct, and conflating them allows each to muddy the other. The sharper question is whether a platform owned by a geopolitical adversary should be operating in European information space at all, and if so, under what structural constraints. That is a harder conversation than “fix the For You feed.”
Historically, the parallel that most illuminates this moment is not media regulation but pharmaceutical regulation: the thalidomide disaster produced the modern drug approval framework, which is genuinely protective but also genuinely slow, conservative, and often captured by incumbent interests. Platform regulation will likely follow a similar trajectory — achieving real protections over decades while also freezing the market structure of the 2020s into place and making it difficult for genuinely novel entrants to challenge incumbents.
What to watch
- Whether the Commission’s enforcement process produces structural algorithm requirements (potentially forcing TikTok to disclose proprietary model details) or settles for user-facing controls.
- The US response: a Democratic Congress may move on its own children’s online safety legislation, potentially harmonising with the EU framework or diverging from it.
- Other major platforms — Meta’s Instagram Reels, YouTube Shorts — and whether they receive comparable scrutiny, or whether TikTok’s Chinese ownership makes it a uniquely convenient target.
- Any legal challenge by TikTok or ByteDance to the DSA’s jurisdictional scope before European courts.
— J