Between accusations of censorship and platform power: What the Digital Services Act actually regulates

Few European laws have triggered as heated a debate in recent times as the Digital Services Act (DSA). Right-wing and right-populist actors on both sides of the Atlantic in particular claim that the DSA suppresses legal opinions, enables digital censorship, and steers online debate. What is regarded within the EU as a step towards greater transparency and accountability for major platforms such as Instagram, Facebook, or TikTok is increasingly attacked in transatlantic discourse as a “censorship law”. Criticism from the United States holds that Europe is restricting freedom of expression online and protecting its own digital economy from the market power of predominantly American tech companies. Within Europe, right-populist voices have taken up this same narrative. Many of these accusations rest on deliberate misrepresentations of the legal situation, others on misunderstandings about the role and reach of EU law. Individual problems in enforcement are generalised, and genuine state censorship – as practised, for instance, in China or Iran – is equated with the rule-of-law application of German and European law. Yet such application is always based on a balancing of different fundamental rights. This piece argues that the DSA, while a complex piece of legislation, has at its core an important purpose: to protect freedom of expression in the digital sphere from the uncontrolled power of private platforms.

When J.D. Vance addressed the Munich Security Conference in February 2025, one of the world’s most important foreign-policy forums, he did not use the platform for conventional security policy. Instead, he warned Europe about digital censorship (Deutschlandfunk, 2025). What he meant was the Digital Services Act, the European law that requires large platforms such as Facebook, TikTok, or X to be more transparent and accountable. His appearance reverberated. The US Congress has since spoken of a “Foreign Censorship Threat” (House Judiciary Committee Republicans, 2025). Politicians from the AfD want to abolish the DSA and accuse the EU of “political persecution instead of the rule of law” (German Bundestag, 2025). Even Tagesschau, Germany’s main news programme, asked: “Is the EU censoring internet platforms?” (Feld, 2025). Elon Musk, whose platform X was fined €120 million by the European Commission in Brussels in December 2025 for failures of transparency, has likewise repeatedly described European regulatory efforts as censorship, called for US sanctions against the EU, and posted that “the EU should be abolished” (Die Zeit, 2025).

The accusation running through all these voices is that ministries, NGOs, fact-checkers, media, academia, and platforms have joined forces in a network to suppress legal opinions, exercise digital censorship, and steer online debate.

Censorship is not just any old term

This criticism resonates because it connects with real experiences. Many people have seen their own posts deleted, their accounts suspended, or their reach curtailed on social media. They perceive that some positions are more socially accepted than others. These perceptions should not be dismissed out of hand. But they do need to be carefully placed in context – otherwise authoritarian practices end up being equated with democratic rule-making.

The legal definition, however, is clear. Under German law, censorship means state pre-publication control of the media aimed at suppressing unwelcome political speech. The Basic Law states unambiguously: “There shall be no censorship” (Article 5(1), third sentence). In Germany, the courts – above all the Federal Constitutional Court – consistently protect freedom of expression and effectively limit state interference. There can therefore be no question of state censorship. Anyone who equates authoritarian practices with the rule-of-law application of European law devalues the concept of censorship, erases Germany’s (and Europe’s) important tradition as a place of free expression, and blinds us to genuine repression.

Whose power the DSA actually limits

Anyone who takes censorship seriously must also take freedom of expression seriously. That means understanding that freedom of expression, unlike, say, the prohibition of torture, is not an absolute right. It is a dynamic right that must be weighed against other protected interests (Kettemann & Müller, 2025). Freedom of expression ends where it touches upon – or even violates – other fundamental rights: hate speech, disinformation, or other people’s rights to dignity and privacy. Article 10(2) of the European Convention on Human Rights states this explicitly: the exercise of freedom of expression “carries with it duties and responsibilities” and may be subject to such formalities, conditions, restrictions, or penalties as are prescribed by law and are necessary in a democratic society in pursuit of a legitimate aim (Council of Europe, 1950, Art. 10(2)). American constitutional law works differently here. The First Amendment prohibits Congress from making laws that abridge “the freedom of speech, or of the press; or the right of the people peaceably to assemble.” “Freedom of speech” is interpreted very broadly in the US, extending even to campaign financing (“commercial speech”) and Holocaust denial.

The DSA’s real innovation lies elsewhere: it shifts the balance of power between platforms and users. For the first time, very large platforms with more than 45 million users in the EU must explain their content-moderation decisions, assess societal risks, and provide effective complaint mechanisms – both on the platform itself and through external dispute-resolution procedures (Mast et al., 2024). If my post is wrongly removed, I now have a clearly defined right to challenge that decision. The DSA makes platform power visible and accountable.

Overblocking: a real problem, but wrongly addressed

One important objection to the DSA deserves serious engagement: platforms delete too much. The term “overblocking” describes this excessive removal of content that is, in fact, legal. It is a real phenomenon, one that affects fundamental freedoms and raises genuine concerns for democratic theory. The same applies to the opposite problem: underblocking – the late or failed removal of genuinely illegal content – is equally problematic (Ohnesorge, 2025).

The cause of both over- and underblocking does not lie with the platforms alone. A legal framework that consistently punishes the failure to remove illegal content more severely than the (accidental) removal of legal content structurally creates an incentive towards excessive caution. This is also a matter for the Digital Services Coordinators responsible for enforcing the DSA: they should signal to platforms that careless overblocking can itself constitute a systemic risk (Watolla et al., 2025), and that careful review is the better path.

Overblocking is thus not evidence of state censorship. It is evidence of the structural power of private platforms – precisely the power the DSA is meant to limit. Research points to ways this could be improved, for instance through proposals for better human–machine interaction in content governance (Kettemann et al., 2025).

Trusted flaggers and fact-checkers: censors?

Another accusation levelled at the DSA is that so-called “trusted flaggers” and fact-checkers are an extended arm of state-directed opinion control. But what do trusted flaggers actually do? In most cases, they are civil-society organisations or public authorities – such as consumer-protection associations or law-enforcement agencies – that can alert platforms to potentially unlawful content. They report; they do not decide on removals. Platforms are obliged to review such reports promptly, not to follow them blindly. The final decision remains with the platforms, and for the first time these decisions are subject to clear obligations regarding transparency, justification, and complaints procedures.

Fact-checks, in turn, have no specific legal effect (and even their practical effect is contested). They contradict false or misleading claims, provide context, or reinforce existing doubts. Calling this censorship equates criticism with prohibition and implicitly elevates one’s own position to an unquestionable truth. Such an understanding of freedom is incompatible with the Basic Law.

That said, it is also clear that state funding of civil-society actors is a sensitive matter and demands transparency and a clear separation of roles. But NGOs do not delete content, suspend accounts, or impose sanctions. They hold no public powers whatsoever. That they advocate normative positions – for example, against racism, antisemitism, or disinformation – is not evidence of suppression but an expression of political contestation within a pluralistic society. A state is free to support NGOs that champion human rights and the protection of minorities rather than those that incite hatred against migrants or people who think differently – and, given the principle of human dignity enshrined in the Basic Law, it may even be obliged to do so.

Double standards

It is also revealing to look at those who raise the loudest accusations of censorship. A pattern emerges in the debate over the DSA and European legislation more broadly: what counts as censorship often depends on whether it conflicts with one’s own opinion. This is particularly clear in the United States. Brendan Carr, chairman of the Federal Communications Commission, publicly pressured Disney to take late-night host Jimmy Kimmel off the air and threatened broadcasters with licence reviews if they continued airing his show (Stelter, 2025). Yet when Grok, the AI chatbot on Elon Musk’s platform X, generated and shared nude images of celebrities against their will, and the British government criticised this, Musk described the UK government as “fascist” (Tagesschau, 2026).

The DSA protects the freedom of public speech

A democratic public sphere depends on argument, criticism, and interpretation. Science distinguishes between better and worse explanations; journalism between substantiated and unsubstantiated claims; civic education between democratically legitimate and illegitimate positions. Anyone who brands these distinctions as censorship fails to recognise that freedom of expression does not exist in a vacuum. It protects the right to speak, not the right to go unchallenged or to compel social approval.

This does not mean the Digital Services Act is above criticism. Precisely because the DSA reaches deep into the institutional architecture of the digital public sphere, its implementation must be monitored with sensitivity to freedom of expression. One central risk is that platforms, fearing sanctions, reputational damage, or regulatory pressure, may act too cautiously and restrict, downrank, or remove more content than is legally required. Regulation can thus unintentionally create incentives for excessive control. The internal risk-management obligations placed on platforms are similarly ambivalent. When very large platforms are required to continuously identify, assess, and mitigate systemic risks, this can increase transparency and accountability. At the same time, there is a danger that societal conflicts are translated into technical compliance processes, in which private companies, external auditors, and regulatory authorities increasingly determine which communicative risks are deemed acceptable. The boundary between legitimate risk management, private content control, and indirect state-driven steering of expectations therefore requires careful ongoing scrutiny.

Freedom of expression is too important to be worn down in a rhetorical barrage against every unwelcome piece of legislation. The Digital Services Act is not a gag order. It is imperfect, as any regulation of complex matters inevitably is. But it represents a serious attempt to protect freedom in the digital sphere from overwhelming power, opacity, and private arbitrariness. The criticism from the United States of Europe’s approach to defending freedom in digital spaces does not hold up. For the most part it is simply inaccurate, and where it does have some basis, it is driven less by a genuine concern for protecting freedom than by foreign-policy calculation. Yet uncritical allegiance to the DSA would serve it just as poorly. Precisely in the year ahead of the DSA review process, critical engagement is needed. Researchers must continue to work towards embedding its application firmly within the rule of law, limiting excessive incentives for control, and ensuring that platform oversight does not itself become a new infrastructure for subtle steering of communication. Anyone who dismisses the DSA wholesale as censorship undermines the very thing they claim to be defending: the freedom of private and public speech.

References

BVerfG, Judgement of 15 January 1958, 1 BvR 400/51, 31. https://www.bundesverfassungsgericht.de/SharedDocs/Entscheidungen/DE/1958/01/rs19580115_1bvr040051.html

Deutscher Bundestag. (2025, 19 December). Fraktionen positionieren sich zum Digital Services Act der Europäischen Union. https://www.bundestag.de/dokumente/textarchiv/2025/kw51-de-aktuelle-stunde-dsa-1133782

Deutschlandfunk. (2025, 14 February). US-Vizepräsident Vance wirft Europa Unterdrückung der Meinungsfreiheit vor. https://www.deutschlandfunk.de/us-vizepraesident-vance-wirft-europa-unterdrueckung-der-meinungsfreiheit-vor-102.html

Die Zeit. (2025, 6 December). Elon Musk fordert Abschaffung der EU. https://www.zeit.de/politik/ausland/2025-12/elon-musk-eu-x-strafe

Europarat. (1950, 4 November). Konvention zum Schutze der Menschenrechte und Grundfreiheiten (Art. 10 Abs. 2).https://www.echr.coe.int/documents/d/echr/convention_DEU

Feld, C. (2025, 25 January). Zensiert die EU die Internet-Plattformen? Tagesschau.https://www.tagesschau.de/ausland/europa/eu-dsa-plattformen-zensur-100.html

House Judiciary Committee Republicans. (2025, 25 July). The foreign censorship threat: How the European Union’s Digital Services Act compels global censorship and infringes on American free speech. https://judiciary.house.gov/media/press-releases/foreign-censorship-threat-how-european-unions-digital-services-act-compels

Kettemann, M. C., Mosene, K., Stenzel, M., Mahlow, P., Pothmann, D. & Spitz, S. (2025). Code of Conduct on Human-Machine Decision-Making in Content Moderation. https://doi.org/10.5281/zenodo.17650987

Kettemann, M. C., & Müller, M. (2025). Das neue Digitalrecht der EU im Praxistest Welche Herausforderungen warten bei der Umsetzung? Friedrich-Ebert-Stiftung, FES Impuls. https://library.fes.de/pdf-files/a-p-b/21536.pdf

Mast, T., Kettemann, M. C. , Dreyer, S., & Schulz, W. (2024). Digital Services Act / Digital Markets Act. Beck. https://www.beck-shop.de/mast-kettemann-dreyer-schulz-dsa-dma/product/34660327 

Ohnesorge, J. (2025). Counting without accountability? An analysis of the DSA’s transparency reports. In Digital Society Blog. Zenodo. https://doi.org/10.5281/zenodo.17201618 

Stelter, B. (2025, 18 September). How Brendan Carr, the attack-dog FCC chair, helped take down Jimmy Kimmel with words, not actions. CNN. https://edition.cnn.com/2025/09/18/media/brendan-carr-jimmy-kimmel-fcc-first-amendment

Tagesschau. (2026, 12 January). Musk attackiert britische Regierung im KI-Streit. Tagesschau. https://www.tagesschau.de/wirtschaft/digitales/musk-grok-ki-grossbritannien-100.html

Watolla, A., Zerrer, P., Rau, J., Merten, L., Kettemann, M. C., & Puschmann, C. (2025). Gesellschaftliche Auswirkungen systemischer Risiken. Demokratische Prozesse im Kontext von Desinformationen (Studie für die Bundesnetzagentur). https://www.dsc.bund.de/DSC/DE/Aktuelles/studien/Auswirkungen%20Systemischer%20Risiken.pdf