Content moderation has become one of the central governance questions of the digital age. Decisions about whether to remove online speech determine the contours of public discourse. Despite growing regulatory intervention, particularly in the European Union, dissatisfaction with the current model persists.

Concerns about the over-removal of lawful speech, a lack of contextual sensitivity, and the concentration of power in private platforms remain unresolved. The challenge is not how to moderate content effectively, but how to do so in a way that is legitimate, proportionate, and compatible with freedom of expression.

In partnership with Analysis & Numbers, The Future of Free Speech has launched a new distributed content-moderation prototype that offers a concrete alternative. Rather than reinforcing centralized moderation, it redistributes decision-making power to users and communities.

The key rationale for the prototype’s development is that such distributed moderation models directly address structural flaws in both platform governance and regulatory approaches and may represent a necessary evolution in content moderation design.

The Structural Limits of Centralized Moderation

Content moderation has been defined as the process by which platforms screen, evaluate, categorize, approve, or remove user-generated content in accordance with relevant policies. Centralized moderation operates through standardization. Platforms apply uniform rules across diverse contexts, often relying on automated systems to manage scale. This produces efficiency but at the cost of nuance.

Moderation decisions are shaped not only by legal requirements but by platform incentives, risk management strategies, and internal governance structures. The European Union’s Digital Services Act (DSA) represents the most comprehensive attempt to regulate content moderation. It imposes obligations on platforms to remove illegal content, implement notice-and-action mechanisms, and ensure transparency and accountability.

However, the DSA operates within, and arguably reinforces, the centralized model it seeks to regulate. It assumes identifiable intermediaries capable of exercising control over content and enforcing rules across their systems. This creates a regulatory paradox. While the DSA aims to protect fundamental rights, including freedom of expression, it simultaneously incentivizes platforms to adopt more restrictive moderation practices.

One consequence of this regulatory framework is systematic over-removal. Platforms are incentivized to err on the side of caution, particularly where regulatory or reputational risks are high. In a 2024 study, we found that a substantial majority (87.5% to 99.7%) of deleted comments on Facebook and YouTube in France, Germany, and Sweden were legally permissible, suggesting that platforms, pages, or channels may be over-removing content to avoid regulatory penalties. This could reflect how regulatory pressure can exacerbate this tendency, encouraging platforms to remove lawful but controversial content in order to avoid liability or public backlash.

Automation intensifies this dynamic. AI-based moderation systems, while necessary at scale, struggle with context, irony, and cultural nuance. Empirical research has shown that such systems can produce both false positives (over-removal) and false negatives (under-enforcement), raising concerns about both accuracy and fairness. The result is a system in which speech is filtered not only by law but also by platform risk calculations embedded in algorithmic design.

Moreover, regulation struggles to keep pace with technological change. Emerging environments such as metaverses challenge the assumption that content can be centrally controlled and removed.

Subscribe now

Decentralization and the Crisis of Control

The rise of decentralized technologies introduces a fundamental shift in the structure of online systems. In decentralized networks, data is distributed across multiple nodes, with no single entity exercising full control. This has significant implications for content moderation.

Traditional models rely on central intermediaries to detect, evaluate, and remove content. In decentralized systems, these functions become fragmented or absent. Decentralization can enhance privacy, security, and resistance to censorship. At the same time, it creates new challenges for governance. Without central control, enforcing rules becomes difficult, and harmful content may persist.

Against this backdrop, distributed moderation offers a different paradigm. Rather than concentrating power in platforms or relying solely on regulation, it redistributes decision-making to users and communities.

The prototype provides a practical example of this approach, integrating AI classification with user-defined moderation thresholds, enabling context-sensitive decision-making. The system operates through a structured pipeline: comments are collected via platform APIs, analyzed by machine-learning models, stored in a database, and presented to moderators via a prioritized interface. Moderators can then take actions such as hiding, deleting, or blocking content.

Crucially, the system does not automate final decisions. AI is used to assist, not replace, human judgment. This hybrid model addresses the limitations of both purely automated and purely manual moderation. Normatively, the system embodies a shift from standardization to pluralism. Different communities can adopt different thresholds for what constitutes harmful content, reflecting their specific contexts and values.

The Pros and Cons

Distributed moderation directly addresses the structural problems of centralization. First, it reduces over-removal by allowing communities to calibrate their own thresholds. This counters the precautionary bias inherent in centralized systems and aligns moderation more closely with legal standards of permissible speech.

Second, it enhances contextual sensitivity. Community moderators are better positioned to understand the nuances of language, culture, and interaction within their spaces. This improves both accuracy and legitimacy.

Third, it restores user agency. As noted in research on decentralized governance, empowering users can “democratize the process” of content regulation and reduce reliance on opaque platform decisions. These features are particularly important given concerns about the concentration of power in digital platforms. By redistributing moderation authority, distributed systems can mitigate the risks associated with centralized control. At the same time, the model avoids the pitfalls of full decentralization. By retaining a structured interface, AI support, and integration with existing platforms, it ensures that moderation remains feasible and effective.

However, distributed moderation is not a panacea. It introduces new challenges, particularly regarding consistency, accountability, and potential fragmentation. Different communities may adopt divergent standards, raising questions about equality and fairness. There is also a risk that some communities may tolerate harmful content, particularly in the absence of strong oversight.

Moreover, empowering users requires capacity. Effective moderation depends not only on tools but on training, resources, and institutional support. Without these, distributed systems may struggle to achieve their full potential.

There are also legal questions. Existing regulatory frameworks are designed for centralized actors, and adapting them to distributed models will require careful consideration. Nevertheless, these challenges are not unique to distributed systems. They reflect broader tensions in content moderation between centralization and pluralism, efficiency and legitimacy, safety and freedom.

Can Decentralization Be The Future?

The current model of content moderation is reaching its limits. Centralized systems struggle with context, scale, and legitimacy, while regulatory frameworks risk reinforcing the very dynamics they seek to correct. Distributed moderation offers a promising alternative that could rebalance the relationship between platforms, users, and regulators.

The prototype discussed in this article demonstrates that such models are not merely theoretical. They can be built, tested, and deployed. The question is not whether distributed moderation will replace centralized systems entirely. It is whether it can complement and reshape them in ways that better align with democratic values and fundamental rights.

As online environments continue to evolve, particularly with the rise of decentralized technologies and immersive platforms, the need for such rethinking will only become more urgent.

Natalie Alkiviadou is a Senior Research Fellow at The Future of Free Speech. Her research interests lie in the freedom of expression, the far-right, hate speech, hate crime, and non-discrimination.

Over the past decade, journalists and civil society groups have raised concerns about media pluralism and press freedom in Hungary. The Hungarian media landscape, observers have noted, has undergone substantial transformation since 2010, with ownership structures and regulatory policies evolving in ways that favor outlets aligned with the government. At the same time, officials have placed increasing pressure on independent media organizations.

Recently, Hungarian regulators denied the renewal of the broadcast license for the prominent independent radio station Klubrádió, thus removing it from the FM airwaves. When Hungary’s Media Council launched a new tender procedure for the same radio frequency, it declared Klubrádió’s application invalid.

Not only did the station’s disappearance from the radio spectrum in Hungary attract significant attention both domestically and internationally, but it also led to a significant legal decision by the Court of Justice of the European Union (CJEU) that demonstrated how seemingly technical and administrative questions can implicate fundamental rights such as freedom of expression.

On 26 February 2026, the CJEU, in Commission v Hungary (Case C-92/23), held that Hungary had breached EU law when its national media regulator refused to renew Klubrádió’s broadcasting license and subsequently excluded the station from a tender procedure for the provision of media services on the same frequency.

The CJEU concluded that the Hungarian authorities had relied on regulatory grounds that were disproportionate and incompatible with EU law and that the resulting measures interfered with the freedom of expression guaranteed by Article 11 of the Charter of Fundamental Rights of the European Union, which protects the freedom of expression and information, including the freedom and pluralism of the media.

The decision, therefore, sits at the intersection of two legal domains that do not always overlap in obvious ways — namely the EU framework governing electronic communications and the protection of fundamental rights, specifically that of expression and information. By doing so, the CJEU reaffirmed that the renewal of licenses and tender procedures cannot be treated solely as administrative matters when they affect media outlets’ ability to reach audiences and participate in democratic debate.

Background on The Dispute

Klubrádió was founded in 1999 and has been described by the European Parliament as “one of the few remaining free voices in the Hungarian media landscape.” In 2014, the Hungarian Media Council granted Klubrádió the right to use the 92.9 MHz frequency in Budapest, allowing it to broadcast in the capital for a period of seven years. The agreement also contained a clause permitting the licence to be renewed once for an additional five-year term under certain conditions. As the license was about to expire, the station applied for renewal in accordance with the applicable legal framework.

The Media Council rejected the application. The authority relied on a provision of Hungarian media legislation according to which a licence could not be renewed if the broadcaster had committed “repeated infringements” of regulatory obligations during the licence period. According to the regulator, Klubrádió had failed to fully comply with certain reporting requirements regarding programming quotas and had therefore committed repeated breaches of national media law.

The infringements cited by the authority were largely administrative in nature. They were related to failures to submit two monthly reports regarding broadcasting obligations. Although the breaches had been identified and sanctioned during the licence period through warnings or modest fines, the Media Council nevertheless considered them sufficient to prevent the renewal of the station’s broadcasting licence.

As a result of the refusal, Klubrádió lost its right to broadcast on the Budapest FM frequency. In February 2021, the station ceased broadcasting on FM radio and continued operating only online. The decision attracted considerable criticism from journalists, civil society organisations, and international observers, many of whom viewed the move as part of a broader pattern of pressure on independent media outlets in Hungary.

Following the refusal to renew the licence, the Media Council launched a new tender procedure for the 92.9 MHz frequency. Klubrádió submitted an application to participate in the competition. However, the regulator subsequently declared the application invalid. The authority cited technical irregularities in the station’s proposed programming schedule and concerns relating to the company’s financial data.

The European Commission subsequently initiated infringement proceedings against Hungary before the CJEU. The Commission argued that the refusal to renew the licence and the handling of the tender procedure violated EU law.

Spectrum Allocation and EU Law

The radio spectrum (the radio-frequency portion of the electromagnetic spectrum) constitutes a limited public resource that must be allocated to ensure efficient use while preserving fair competition and pluralism in the media market. Although Member States retain primary responsibility for managing spectrum resources, EU law establishes a framework governing how these powers must be exercised.

The EU electronic communications regime requires national authorities to allocate and manage radio frequencies in accordance with principles of objectivity, transparency, non-discrimination, and proportionality. These principles aim to ensure that regulatory decisions affecting access to spectrum do not distort competition or undermine the functioning of the internal market.

In the present case, the Commission argued that Hungary’s regulatory framework and the decisions adopted by the Media Council were incompatible with these requirements. In particular, the Commission maintained that the Hungarian rules allowed relatively minor administrative breaches to automatically lead to the refusal of licence renewal, thereby producing consequences that were disproportionate to the seriousness of the underlying infringements.

The Commission also challenged the way the tender procedure for the frequency was conducted. It argued that the reasons given for excluding Klubrádió from the competition were excessively formalistic and inconsistent with the principles governing spectrum allocation under EU law.

Crucially, the Commission further argued that the Hungarian measures interfered with the freedom of expression, media freedom, and pluralism. Because radio frequencies are limited and access to them determines whether broadcasters can reach large audiences, regulatory decisions concerning spectrum allocation may directly affect the functioning of the media landscape. In this sense, the denial of a broadcasting licence may have consequences comparable to restrictions on publication or distribution.

The CJEU’s Reasoning

In its judgment, the CJEU emphasised that while Member States enjoy a degree of discretion in managing spectrum resources, discretion must be exercised in compliance with EU law and the fundamental rights guaranteed by the Charter.

A central issue concerned the Hungarian legal provision that prevented the renewal of a broadcasting licence where the broadcaster had committed repeated regulatory infringements. The CJEU observed that the infringements attributed to Klubrádió were relatively minor administrative breaches relating to reporting obligations. These breaches had already been sanctioned by the competent authority.

Nevertheless, the Hungarian framework allowed those breaches to serve as the decisive reason for refusing to renew the broadcasting license. The CJEU considered that such an outcome failed to respect the principle of proportionality. As the CJEU noted, the regulatory system allowed administrative irregularities to lead to excessively severe consequences when viewed in relation to the nature of the underlying conduct.

The CJEU therefore concluded that Hungary had failed to fulfil its obligations under EU law. The decision to deny the licence renewal was not compatible with the principles governing the allocation of radio spectrum within the EU legal framework.

The CJEU also examined the tender procedure organised by the Media Council after the refusal to renew the licence. Although national regulators may impose technical requirements on applicants, those requirements must be applied in a transparent and proportionate manner.

In this case, the Hungarian authorities rejected Klubrádió’s application due to alleged errors in the station’s proposed programming schedule and questions about its financial position. However, the CJEU concluded that the manner in which the tender procedure was conducted did not comply with EU rules governing spectrum allocation.

The CJEU emphasised that regulatory requirements must not be applied in ways that effectively exclude broadcasters from access to the market without adequate justification. Where technical or administrative irregularities are invoked to invalidate an application, the consequences must remain proportionate to the nature of the irregularities.

Significantly, the CJEU linked its analysis to the protection of freedom of expression and media pluralism under the Charter. The judgment recognised that decisions concerning access to broadcasting frequencies can have a direct impact on a media organization’s ability to operate and participate in public debate.

The CJEU therefore concluded that the Hungarian measures also infringed Article 11 of the Charter. The ruling confirmed that actions that may appear merely administrative cannot be separated from the broader constitutional framework governing fundamental rights within the EU.

Why The Judgment Matters

The importance of Commission v Hungary (Case C-92/23) lies not only in its outcome but also in the reasoning adopted by the CJEU. The judgment illustrates how EU law governing electronic communications can intersect with the protection of fundamental rights when regulatory decisions affect the media.

The CJEU treated access to broadcasting frequencies as a factor that can influence the effective exercise of freedom of expression, media independence and pluralism. Because radio frequencies are limited and access to them determines whether broadcasters can reach a wide audience, decisions denying such access may significantly affect the diversity of voices available in the public sphere.

By grounding part of its reasoning in Article 11 of the Charter, the CJEU reaffirmed that the freedom and pluralism of the media constitute core values within the EU legal order. The judgment, therefore, signals that national regulatory frameworks governing broadcasting licences must be applied in ways that respect those values.

In practical terms, the ruling also sends a clear message to national regulators across the EU. Administrative compliance rules and technical requirements must not be used in ways that effectively exclude media outlets from the market through disproportionate enforcement measures. Where regulatory decisions restrict access to broadcasting frequencies, authorities must ensure that the resulting consequences remain proportionate and justified.

Conclusion

The CJEU’s judgment in Commission v Hungary (Case C-92/23) demonstrates how regulatory disputes that appear technical on the surface can raise profound questions about democratic governance and fundamental rights. What began as a disagreement over the renewal of a broadcasting licence ultimately became a test of whether national spectrum management rules could be applied in ways that undermine media pluralism.

By concluding that Hungary had breached EU law and violated the Charter’s protection of freedom of expression, the CJEU reaffirmed the importance of proportionality, transparency, and non-discrimination in spectrum regulation. The ruling also highlights the increasingly important role that EU law can play in safeguarding media freedom when national regulatory systems operate in ways that disadvantage independent outlets.

Whether the judgment will lead to broader reforms in Hungary’s media licensing system remains uncertain. What is clear, however, is that the CJEU has established an important precedent. The allocation of broadcasting frequencies cannot be treated solely as a technical-administrative matter when the consequences of regulatory decisions affect media organizations’ ability to operate and participate in democratic debate. Where those decisions threaten the freedom and pluralism of the media, EU law may intervene to ensure that fundamental rights are respected.

Natalie Alkiviadou is a Senior Research Fellow at The Future of Free Speech. Her research interests lie in the freedom of expression, the far-right, hate speech, hate crime, and non-discrimination.

Subscribe now

Share

In a decision this month, the European Court of Human Rights (ECtHR/Court) offered a rare moment of doctrinal clarity in its historically uneasy relationship with provocative expression under Article 10 of the European Convention on Human Rights. Even more surprising is that the speech in question involved a satirical comparison to Nazi symbols and figures.

The judgment in Mladina d.d. Ljubljana v Slovenia (No.2) (2026) stands out not because it revolutionizes Article 10 jurisprudence, but because it faithfully applies principles the Court has long articulated but has routinely abandoned when expression becomes uncomfortable.

When read alongside other cases involving the use of totalitarian symbols, the ruling highlights the Court’s inconsistent and incoherent application of context-based analysis to historically charged imagery and language.

The Mladina Ruling

The ECtHR judgement arose from defamation proceedings against Mladina, a Slovenian political weekly known for its satirical section. At the time, the claimant, B.G., was a prominent opposition member of parliament for the Slovenian Democratic Party (SDS).

The controversy followed a public debate sparked by a Facebook post comparing B.G. to Nazi propaganda minister Joseph Goebbels. Responding to that debate, Mladina published an editorial criticizing what it described as double standards in Slovenian journalism, drawing parallels between the political communication methods of the SDS and Nazi propaganda.

In the same issue, Mladinamit (the satirical section of the weekly) published a satirical piece mocking the comparison. Beneath it, two photographs were placed side by side, one of Goebbels with his family and the other of B.G. with his family, presented in an identical format. The photo of B.G. and his family had been taken at a large public religious event and had previously appeared in other media. B.G. subsequently brought defamation proceedings against the magazine.

Slovenian courts accepted that the textual criticism constituted protected political speech, but held that the photographic comparison crossed the line. Images, they reasoned, intruded more deeply into private life than text, and placing the politician’s family, particularly children, next to Goebbels’ family amounted to an unjustified attack on reputation. The magazine was ordered to publish an apology, publish the judgment, and pay damages.

The ECtHR disagreed and found a violation of Article 10. Central to the Court’s reasoning was the finding that the publication “contributed to a debate of general interest,” namely, the permissibility of comparing contemporary political methods with those of the Nazi era. Stressing that satire is a form of expression which “by its inherent features of exaggeration and distortion of reality, naturally aims to provoke and agitate,” the Court found that the domestic courts had failed to strike a fair balance between Article 10 and Article 8 (the right to private and family life).

While the juxtaposition of family photographs could engage the right to private life, the Court emphasized that the publication appeared in a clearly satirical section, concerned a well-known politician who had himself exposed his family to public attention, and was addressed to a limited and politically literate readership.

Was The Ruling Too Context-Specific?

Here, unlike in previous cases, the Court takes context, genre, and political function seriously, recognizing satire as a form of expression that operates precisely through exaggeration, discomfort, and symbolic provocation.

There are still some elements of this decision worthy of criticism. The Court assumes that readers of the magazine’s satirical section would readily understand the political context and message conveyed by the photographs, echoing the problematic audience-based reasoning seen in Féret v Belgium (2010).

In Féret, the Court justified restricting anti-immigrant, anti-Islam statements made by a politician and incorporated into publications of his party. In its judgment, the ECtHR noted that the statements were “inevitably of such a nature as to arouse, particularly among the less informed members of the public, feelings of distrust, rejection or hatred towards foreigners.” This premise was sharply criticized in a dissenting opinion for portraying members of the public as “nitwits… incapable of replying to arguments and counter-arguments, due to the irresistible drive of their irrational emotions.”

But this reasoning raises an uncomfortable question. Would the expression in Mladina have received the same level of protection had it appeared in a tabloid newspaper or been circulated randomly online? The Court’s ruling suggests that the answer may depend less on the content of the speech than on an abstract and deeply contestable assessment of the audience presumed to consume it.

The Court also highlighted the limited impact of the publication, noted the absence of any demonstrated serious harm to the politician’s own reputation, and recalled that “comparisons to the German Nazi regime do not automatically justify a conviction for defamation on the ground of the special stigma attached to the latter, especially if there exist special circumstances justifying such a comparison.”

The stigma attached to Nazism thus operates as a quasi-exclusionary category, shaping the analysis before balancing has meaningfully begun. This framing grants courts wide discretionary power to determine, ex post facto, what speech it finds tolerable. The Court presents “special circumstances” as factors capable of preventing an automatic restriction on Nazi comparisons.

In principle, this appears protective of Article 10, since it rejects the idea that such comparisons are per se defamatory. Yet the reasoning's structure is problematic. By foregrounding the “special stigma” of Nazism as the starting point, the Court implicitly treats Nazi comparisons as presumptively suspect, requiring contextual justification to escape sanction. “Special circumstances” are thus corrective as they may operate to neutralise a stigma that the Court has already elevated into a normative baseline.

Furthermore, the ECtHR rejected the domestic courts’ fragmentation of text and image, recalling that “the form of expression cannot be dissociated from its context and apparent goal.” This is a simple but often neglected position that satire must be assessed as satire. Readers do not encounter satirical publications by dissecting images and text in isolation. Satire is a composite form, one that Mladina recognizes. This contributed to the Court’s resistance to sanitizing political critique under the guise of protecting dignity.

Context, Satire, and the ECtHR’s Approach to Totalitarian Imagery

Satire relies on distortion and provocation to challenge power and dominant narratives. Its value lies precisely in its ability to unsettle. The ECtHR’s approach in Mladina aligns with its better satire jurisprudence, though that jurisprudence has not always been applied consistently. The Court has previously acknowledged that satire “aims to provoke and agitate,” and that its exaggerated nature must be taken into account when assessing alleged harm.

Mladina avoids this pitfall. Rather than asking whether the Nazi comparison was offensive in the abstract, the Court examined why the comparison was made and how it functioned within a broader political critique. This is consistent with the core Article 10 principle, that freedom of expression protects not only information or ideas that are favourably received, but also those that “offend, shock or disturb.”

Although Mladina was not framed as a hate speech case or as one concerning unconstitutional symbols, the use of Nazi figures and the Court’s explicit references to the German Nazi regime make comparison with its totalitarian-symbols case law both relevant and instructive.

The contrast between Mladina and Nix v Germany (2018) is stark. In the latter, the applicant used his personal blog to criticize the treatment of his German-Nepalese daughter by a State employment office, situating his posts within a broader public debate on discrimination against children from migrant backgrounds and families on social welfare. In one post, he illustrated his critique by reproducing a photograph of Heinrich Himmler in SS uniform bearing a swastika, alongside a historical quotation on Nazi education policy. He was convicted under German criminal law for using symbols of unconstitutional organizations (his libel charges were dropped at the domestic level). The ECtHR declared his Article 10 complaint inadmissible.

In Nix, the ECtHR emphasized that states marked by the legacy of Nazism, because of their history, “may be regarded as having a special moral responsibility to distance themselves from the mass atrocities perpetrated by the Nazis.” In doing so, it shifted the analysis from whether bans are permissible under Article 10 to an implicit suggestion that they may be obligatory. In this framework, the Court noted that the symbol in question, a photograph of Heinrich Himmler in SS uniform prominently displaying a swastika armband, “cannot be considered to have any other meaning than that of Nazi ideology.” The courts characterized the Nazi imagery as no more than an “eye-catching device,” devoid of any meaningful link to the substance of the blog post.

Accepting this reasoning, the ECtHR concluded that Nix’s use of SS symbols, figures, and uniform constituted a “gratuitous use of symbols,” which domestic law was entitled to prohibit. Mladina is significant precisely because it resists the prohibitionist reflex that marks Nix. What mattered in Mladina was whether the comparison was anchored in political critique and assessed in context. By contrast, in Nix, Nazi symbolism was treated as having a fixed, monolithic meaning that overrides intent, genre, and discursive function.

Importantly, the ECtHR has not symmetrically applied historical sensitivity as was the case in Nix. For example, in Behar and Gutman v Bulgaria (2021), which involved anti-Semitic speech by a politician, the Court noted that “the specific historical context in respect of Bulgaria, as opposed to that in respect of other European States, matters little in that regard.” In other words, the absence of a pronounced domestic history of anti-Semitism did not dilute the Court’s condemnation of such expression.

In Vajnai v Hungary (2008), which involved the prohibition of the wearing of the communist red star at a demonstration, the Court recognized the “terror” and “scars” left by communist regimes in countries such as Hungary, yet held that these emotional legacies, on their own, were insufficient to justify restrictions on Article 10.

Fáber v Hungary (2010) involved the prohibition of the striped Árpád flag, a medieval symbol later appropriated by Hungary’s fascist Arrow Cross Party, near two separate demonstrations. In finding a violation of Article 10, the ECtHR highlighted that “ill feelings or even outrage, in the absence of intimidation” could not justify restrictions on expression. A dissenting opinion emphasized that, in the Court’s earlier case law on Holocaust denial and related expression, both the Court and the Commission had consistently denied Convention protection to such speech.

Regrettably, the Court provided no detailed justification for this conclusion and failed to explain why a symbol associated with the Arrow Cross regime, whose ideology closely mirrored Nazism and which was responsible for the mass murder of Jews in Hungary, was treated differently from Nazi symbols. The question, of course, remains: had it been an actual Nazi flag or a Schutzstaffel flag, would the Court have reacted differently?

What Article 10 Could Look Like If Applied Consistently

Mladina matters because it shows that the Court can robustly apply Article 10, even in potentially controversial cases. It demonstrates that protecting political satire is compatible with historical sensitivity, that Nazi comparisons are not per se abusive, and that balancing exercises must focus on context, function, and risk rather than on symbolic absolutism. Mladina’s methodology should not be exceptional. The principles applied, holistic reading, genre sensitivity, and rejection of speculative harm, are not new. They are simply too often forgotten when expression becomes unsettling.

Freedom of expression is tested at its edges. Mladina reminds us what Article 10 looks like when the Court resists the temptation to retreat from its own standards. The question is whether it will do so consistently. The positive assessment of Mladina should nevertheless be treated with caution. In Nix, Germany’s Nazi past operated not merely as a contextual factor but as a justificatory engine for restriction, allowing historical exceptionalism to do all the work. In Behar and Gutman, the absence of a significant historical connection to Nazism in Bulgaria, compared to other European States, played no role in the Court’s assessment, reflecting the seriousness with which it treats anti-Semitic expression irrespective of national context.

While the Slovenian judgment marks a clear shift away from such exceptionalism, it leaves unresolved challenges, particularly the Court’s assumptions about the interpretive capacities of readers and the fact that the Nazi figures in question appeared in a non-German publication. Whether Mladina signals a lasting recalibration or remains an exception to the Court’s inconsistent approach to controversial speech remains to be seen.

Natalie Alkiviadou is a Senior Research Fellow at The Future of Free Speech. Her research interests lie in the freedom of expression, the far-right, hate speech, hate crime, and non-discrimination.

Share

Subscribe now

Valeria Ilareva, Lidia Staykova, and Krasimir Kanev are human rights defenders in Bulgaria. Ilareva worked as a lawyer in the field of immigration, Staykova volunteered with the State Agency for Refugees, and Kanev headed the Bulgarian Helsinki Committee.

In January of 2015, they were subjected to a barrage of Facebook posts that included explicit death threats and incitement to violence. One user wrote: “Damn, how these monsters mock us? … I feel like shooting them, there are loads of vermin to eliminate.” Another wrote: “Die, carrion. Cyanide for you and for all traitors to the nation.” A third posted: “I say hang them at Parliament, exhibit them there as Christmas decorations so that their bones hang there forever.” Others suggested skinning the applicants alive or smearing them with honey and leaving them to be devoured in an ants’ nest. The messages spread quickly. A photomontage depicting the applicants as “Freaks of the Year” was shared over one hundred and twenty times in four days.

The threats were not abstract.

Ilareva explained that her office address was easily searchable online, while Kanev was later physically attacked in Sofia after appearing on television to criticise anti-immigrant rhetoric. They also filed complaints with the prosecutor’s office, arguing that the posts amounted to death threats, incitement to violence, and hate speech.

Yet prosecutors opened proceedings only under Article 162 of the Bulgarian Criminal Code, the hate speech provision, and declined to pursue the more obvious charge of threats under Article 144 of the Code. Investigations quickly stalled, ostensibly because IP addresses could not be obtained, and prosecutors dismissed the comments as “negative assessments” that could not cause “a justified fear” because they had been made on the internet.

Importantly, in relation to Article 162, national authorities found that the applicants could not fall within the sphere of this article as they did not qualify as victims, which would be the minority groups themselves. When Ilareva, Staykova, and Kanev found no recourse with Bulgarian authorities, they brought their case before the European Court of Human Rights (ECtHR/Court/Strasbourg).

But the ECtHR did not just examine the application (or lack thereof) of the Criminal Code as it relates to death threats. Instead, the Court’s judgment in Ilareva and Others v. Bulgaria (2025) further stretched hate speech doctrine as defined by the ECtHR in ways that will create a confusing hierarchy of harms and undermine clear and consistent law when it comes to freedom of expression.

The ECtHR’s Judgement

This debate sits within a broader human rights framework. Article 4 of the International Convention on the Elimination of All Forms of Racial Discrimination and Article 20 of the International Covenant on Civil and Political Rights specifically prohibit certain types of speech.

The European Convention on Human Rights (ECHR) takes a different approach: Article 10 protects the right to freedom of expression, subject to limitations such as safeguarding the rights of others, but it does not impose an explicit duty on States to restrict speech. Over time, however, the ECtHR has developed such duties through its jurisprudence.

Victims of alleged hate speech often bring claims under Article 8, the right to private and family life, read together with Article 14, which prohibits discrimination. In almost systematically finding in favour of applicants, the result has been a creeping doctrine of positive obligations, where States are expected to investigate and respond to hate speech as part of their Convention duties.

In Ilareva, the applicants relied on Articles 3 (the prohibition of torture, inhuman or degrading treatment or punishment) and 8 of the Convention, alone or in conjunction with Articles 13 (right to an effective remedy) and 14 of the Convention.

The ECtHR unanimously found a violation of Article 8 in conjunction with Article 14. The Court stressed that States’ obligations to protect fundamental rights apply online as much as offline. It criticised the Bulgarian authorities for narrowing the scope of the investigation, refusing to treat the applicants as victims of hate speech, and downplaying the seriousness of the threats.

As such, instead of grounding the violation firmly in Article 8, the Court layered its analysis in Article 14 and constructed a narrative of hate speech by association.

Subscribe now

What Is Hate Speech by Association?

Here is the crux of the problem: the Court did not say Ilareva and her colleagues were victims of hate speech because they were part of a vulnerable group. Instead, it said they were victims because they helped vulnerable groups.

In the Court’s words, the threats and incitement they faced were tied to “their association, through their professional activities, with the groups of people for whose rights they worked.” Specifically, the Court noted that:

“the threats, incitement to violence and hate speech, motivated by intolerance and prejudice and directed against the applicants because of their association, through their professional activities, with the groups of people for whose rights they worked, remained virtually without legal consequences, and the applicants were not provided with the required protection of their right to personal integrity.”

That reasoning is part of a broader pattern. The Court has slowly been expanding who can be considered a victim of hate speech.

In Aksu v Turkey (2012), the Court recognised that negative Roma stereotypes could fall under Article 8 but ultimately found no violation of this right, emphasising the academic nature of the publications in question. In Beizaras and Levickas v Lithuania (2020), the Court confronted homophobic Facebook comments such as “into the gas chamber with the pair of them” and found a violation of Articles 8 and 14 when authorities failed to investigate. In Behar and Gutman v Bulgaria (2021) and Budinova and Chaprazov v Bulgaria (2021), the Court extended protection to communities as a whole, finding violations of Article 8 in conjunction with Article 14 in response to anti-Semitic and anti-Roma rhetoric by a far-right politician, even though there were no direct targets of this speech.

Most recently, in Minasyan and Others v Armenia (2025), the Court held that activists could be victims of hate speech because they were described in a newspaper as “enemies of the nation” for supporting LGBT rights. The case did not involve threats of violence but reputational attacks and calls for exclusion. The Court nonetheless found violations of Articles 8 and 14, reasoning that the applicants were targeted because of their perceived association with the LGBT community. As I noted in a previous piece, this judgment demonstrates “a clear broadening of the Court’s understanding of who can be affected by hate speech.”

Taken together, these cases show a steady expansion of the Court’s approach. In Aksu, the Court acknowledged that stereotyping could be harmful but still maintained the right to freedom of expression. By Minasyan and now Ilareva, that principle has stretched into an open-ended doctrine. Even death threats, which by their nature may fall outside the protection of Article 10, can now be reframed as hate speech when directed at activists because of their associative role.

The grave danger of the move in Ilareva and Others is that it dramatically enlarges the reach of the hate speech doctrine as applied by the Court. By extending hate speech protection to those associated with vulnerable groups, the Court risks eroding the distinction between threats directed at individuals and hate speech directed to the groups themselves.

In this formulation, almost any hostile remark linked to prejudice may be subsumed into the hate speech framework, leaving the category overly broad and imprecise. Moreover, applicants to the court are reclassified as victims of hate speech by association because of their professional or political work.

Suppose an LGBTQ + ally is told, “You deserve to be killed for supporting gay rights.” The gravity of that expression lies in the explicit death threat. But under the Court’s approach, it could instead be reframed as hate speech by association, because the activist is targeted for their solidarity with a vulnerable community. This not only creates conceptual confusion but could also construct a path to a hierarchy of an inherently abominable and unprotected act, that of a death threat.

Equally troubling is the reverse scenario: what if an individual receives death threats because of anti-immigrant or anti-LGBTQ+ speech, expression which under the Court’s case law falls outside Article 10 protection? Does the unprotected nature of the underlying speech make a death threat any less grave, or does it instead highlight the risks of viewpoint-based restrictions?

Conclusion

The speech in question was neither just offensive nor a vague expression of hostility. It included explicit threats of lethal violence against named individuals, disseminated widely, and accompanied by tangible risks given the public visibility of the applicants. In other words, this speech was categorically outside the protection of freedom of expression under international human rights law.

The very problem of the case on a national level was that the authorities did not examine the crime related to death threats, but instead, focused on hate speech. The ECtHR’s focus should have been on that point. Was the expression not severe enough to fulfill the requirements of the relevant provision of the Criminal Code on death threats? If not, why not? If so, then it should have conducted an assessment under Article 8.

The Court did that, but did not stop at Article 8. Instead, it conducted its privacy analysis in conjunction with Article 14’s non-discrimination prohibition. And it is that blend — the 8-14 combined analysis — that raises the real difficulty in this case.

Moreover, had the applicants been subjected to death threats on account of their own protected characteristics, such as sexual orientation or ethnicity, the matter would have fallen squarely within the very rationale for the existence of hate speech laws. Extending protection to those targeted due to their association with protected groups risks diluting the normative clarity and purpose of such laws, without a compelling justification. What required proper investigation in this case was the sheer gravity of a death threat as such, not the alleged or speculative motives underlying it.

The danger is that hate speech becomes a catch-all, applied wherever prejudice is allegedly in the background. This approach obscures the standalone severity of death threats and expands hate speech and Article 14 into new terrain. The Court’s message should have been clear: harassment of activists is not to be tolerated under any circumstances, no matter the reason.

From Aksu to Beizaras and Levickas, from Budinova and Chaprazov to Minasyan, and now in Ilareva, Strasbourg has steadily expanded its positive obligations framework. States have an obligation to investigate hate speech even when the direct victims of hate speech are not the Court’s applicants. While the threats in this case should have been investigated as such (and not as hate speech), recent case law taken together demonstrates that Strasbourg pushes the boundaries of positive obligations in relation to hate speech. The result is that the stakes are getting higher.

The Court needs to establish a clear and consistent framework to justify this expansion. Without it, its uneven reasoning and judicial overreach risks not only weakening protection for vulnerable groups but also chipping away at the core of free expression.

Natalie Alkiviadou is a Senior Research Fellow at The Future of Free Speech. Her research interests lie in the freedom of expression, the far-right, hate speech, hate crime, and non-discrimination.

In July 2025, the Romanian President Nicușor Dan took the extraordinary step of challenging a newly adopted hate speech law before the Constitutional Court. Amongst other issues, the law, passed by the Senate in May, amended existing legislation to enact harsher penalties for antisemitic and xenophobic speech, with even stricter sanctions when such expression is disseminated online.

Romania’s president raised concerns over the law’s vague language and its potential to infringe on freedom of expression. But what makes his challenge unique is that Romania’s law aligns seamlessly with prevailing European standards. It also contributed to the complexity of what he was attempting. A few days after his initiative, the Court struck down his case.

Both the European Court of Human Rights (ECtHR) and the European Union’s Framework Decision on Racism and Xenophobia not only permit such sweeping speech restrictions; they require them.

Subscribe now

Legal Standards Under the European Convention

Under Article 10 of the European Convention on Human Rights (ECHR), freedom of expression may be restricted only if the limitation is prescribed by law, pursues a legitimate aim, and is necessary in a democratic society. In its landmark judgment in Handyside v The United Kingdom (1976), the Court famously stated that freedom of expression protects not only information or ideas that are “favourably received” but also those that “offend, shock or disturb.”

Yet, in practice, the ECtHR has systematically failed to uphold these standards in hate speech cases. In fact, the ECtHR has moved toward what I call a “low-threshold hatred paradigm,” allowing governments to criminalize speech that merely insults or ridicules certain protected groups. This jurisprudential shift undermines Handyside and signals a weakening of the Court's commitment to freedom of expression.

In relation to xenophobia, one of the most telling examples of the ECtHR’s restrictive approach to speech is Féret v Belgium (2009), where a Belgian politician was convicted for distributing anti-immigrant leaflets during an election campaign. The ECtHR upheld the conviction without requiring any link between the speech and a specific act of violence or discrimination. Instead, it held that:

“incitement to hatred does not necessarily require a call for acts of violence or other unlawful acts. Attacks on individuals through insults, ridicule or defamation targeting certain segments of the population may suffice for authorities to prioritise combating racist speech over the freedom of expression when exercised irresponsibly.”

The ruling, which imposed criminal penalties, set a dangerous precedent for restricting political discourse. This was followed in Le Pen v France (2010) and Zemmour v France (2022). Zemmour highlights the reality that, more than a decade after Féret and Le Pen, the Court has continued to apply the same restrictive standard for free speech regarding xenophobic speech, without further justification or assessment of the impact of this approach.

Nor does it consider that, despite restrictive approaches to even prejudicial speech, the phenomenon of hate speech in Europe has been on the rise in tandem with the growing presence of far-right groups in France, and in Europe more generally. Furthermore, the low threshold hatred paradigm persists in cases involving politicians, despite the Court's acknowledgment that political speech is fundamental to a democratic society and should only be subject to narrow restrictions under Article 10(2). As such, its professed strong protection of freedom of expression for politicians appears difficult to align with its rulings in Féret, Le Pen, and Zemmour.

When it comes to anti-Semitism, the other element of the Romanian law relevant to free speech, the situation is even more dire. The ECtHR has consistently upheld restrictions on Holocaust denial on the grounds that it is a primary driver of anti-Semitism. Relevant cases are frequently dealt with under Article 17 of the ECHR, which prohibits the abuse of Convention rights.

Applying this article means bypassing the legality, necessity, and proportionality tests under the limitation grounds of Article 10(2). For example, in Garaudy v. France (2003), Williamson v. Germany (2019), and Pastörs v. Germany (2019), the Court refused to apply the legal tests under the limitation grounds of Article 10, instead finding that Holocaust denial and revisionism could not be considered protected speech under Article 17. These rulings have rendered vast categories of expression presumptively punishable and are normatively founded on the case of Lehideux and Isorni (1998). There, the ECtHR stipulated that there exists a “category of clearly established historical facts – such as the Holocaust – whose negation or revision would be removed from the protection of Article 10 by Article 17.”

The ECtHR adopted a markedly different approach to the denial of the Armenian genocide, revealing an inconsistency in its treatment of historical events. While Holocaust denial cases are typically excluded from protection under Article 10 via Article 17 of the Convention —on the grounds that such speech promotes antisemitism and undermines democracy —the Court in Perinçek v. Switzerland (2015) found that public statements denying the Armenian genocide were protected speech. The Court argued that the Armenian genocide remains a subject of historical and legal debate and therefore falls within the scope of legitimate public discourse.

The Court justified this distinction by referencing the temporal and geographic distance of the events, the absence of a legal consensus on the Armenian genocide across Europe, and the fact that the Holocaust was judicially confirmed by the Nuremberg Tribunal. However, this reasoning has been criticised for reinforcing a hierarchy of suffering and denying equal dignity to victims of other genocides. Dissenting judges in Perinçek pointed out that:

“the suffering of an Armenian because of the genocidal policy of the Ottoman Empire are not worth less than those of a Jew under the Nazi genocidal policy. And the denial of Hayots Tseghaspanutyun...or Meds Yeghern...is not less dangerous than Holocaust denial.”

While the horrors of the Holocaust demand ongoing remembrance and moral reckoning, the approach taken by the ECtHR, and mirrored in Romania’s new law, raises serious concerns for free speech. By relying on Article 17 to exclude Holocaust denial from protection altogether, the Court bypasses the careful legal balancing normally required under Article 10(2).

This move, however understandable in emotional and historical terms, creates a precedent where entire categories of expression are removed from democratic scrutiny. The risk is not only that legitimate dissent or historical inquiry may be caught in the net, but also that states are encouraged to expand these carve-outs beyond the Holocaust, applying them to speech that is controversial but not hateful. Shielding memory through law must not come at the cost of silencing debate. In addition, the above-described differentiation in treatment of the Armenian genocide effectively creates an implicit hierarchy of historical suffering, diminishing the recognition afforded to Armenian genocide victims compared to Holocaust victims.

The European Union

The European Union’s Framework Decision on Combatting Certain Forms of Expressions of Racism and Xenophobia by Means of Criminal Law provides that Member States must criminalize certain forms of hate speech, including the denial or trivialization of genocide. While the Decision nods toward fundamental rights and includes some safety nets, such as public order as a necessity to instigate criminal action, it ultimately hands governments broad discretion to police speech with minimal checks. When examining the Framework Decision, as well as ECtHR case law, it becomes clear that there is no regional forum that provides robust protection of the freedom of expression. National laws like Romania’s are not anomalies but the natural consequence of European jurisprudence and law.

The Broader Question

The broader risk is that these speech laws, while ostensibly targeting hate, have a chilling effect on democratic discourse. When states criminalize ideas based on emotional impact or public discomfort, rather than incitement or physical harm, it is dissent that suffers. This is not merely theoretical. For example, in Hungary and Bulgaria, anti-LGBTQ+ speech laws framed as child protection measures have stifled discussion and education.

In addition, there appears to be no assessment of the harms of hate speech laws by European lawmakers or judges. While hate speech can and does result in harm (the level of harm is disputed), the response does not necessarily have to be regulation in all cases. Proponents of regulation, such as Waldron, argue that it safeguards inclusiveness and equality, while critics like Strossen and Rauch warn of overreach, potential misuse, and the stifling of dissent.

In the digital age, challenges multiply, as regulations may inadvertently push harmful speech into unregulated spaces, creating echo chambers that reinforce extremist views. As Shaw aptly observes, cyberspace and those who participate in it make perpetrators “more likely to become entrenched in their hateful beliefs when given the legitimacy of a global audience.” This insight is increasingly applicable to the digital age, where attempts to silence hate speech may inadvertently amplify its impact and further radicalize those spreading it. Thus, any response to hate speech must carefully balance its harm with the unintended consequences of regulation, particularly in the complex and evolving digital environment.

Speech should only be restricted when it has the demonstrable potential to incite real-world violence, as determined by factors such as context, timing, and typology. Regulation risks becoming overly expansive, leading to unintended consequences such as the suppression of dissent, the amplification of extremist rhetoric, and the erosion of public trust in the legitimacy of regulation. Rather than adopting a “plaster approach” that seeks to regulate speech indiscriminately, the focus should be on addressing the structural inequalities that underlie and perpetuate the marginalisation of vulnerable groups. These deeper systemic issues are the root cause of hate speech’s most damaging effects and require long-term, transformative solutions that empower marginalised voices to participate fully and equitably in public discourse. Hate speech regulations can and have often targeted the very marginalized voices they are intended to protect.

In light of the above, what is needed is a recalibration of legal standards at the European level. The stakes are not small. When vague concepts like “xenophobia” become punishable crimes, and when these crimes carry heavy prison sentences, democracies erase the line between hate and disagreement. And when both Strasbourg and Brussels affirm such laws, they become complicit in undermining the very rights they claim to protect.

In this landscape, the Romanian president’s referral of the law to Romania’s Constitutional Court is not just a domestic procedural move. It is one of the few gestures toward accountability in a continent where the legal consensus is tilting against speech. Although his steps were struck down by the domestic court, this could serve as a paradigm for others to follow. Unless the current European approach is challenged, the laws and courts of Europe will continue to exert a quiet but profound chilling effect.

Share

Natalie Alkiviadou is a Senior Research Fellow at The Future of Free Speech. Her research interests lie in the freedom of expression, the far-right, hate speech, hate crime, and non-discrimination.

Should democracies be able to ban or ostracize political parties with “extremist” or “anti-democratic” positions? While our instinct might be to try to oust such views from society lest they spread, propagate, and undermine democratic institutions, both human rights law and history offer valuable lessons for why this approach backfires.

In 2013, economists and scholars opposed to Germany’s involvement in Eurozone bailouts during the sovereign debt crisis founded the Alternative für Deutschland (AfD). Since then, the party’s platform has changed dramatically from conservative Euroscepticism to far-right populism characterized by anti-Muslim, anti-immigration, and nationalist rhetoric. Its leadership has repeatedly flirted with ethnonationalism and historical revisionism.

Despite, or perhaps because of, its radical stances, the AfD has experienced rising electoral success. In the 2025 federal elections, the party won approximately 20.8% of the national vote, making it the second-largest political force. This significantly increased from its 10.4% share in the 2021 elections.

Germany’s domestic intelligence agency, the Federal Office for the Protection of the Constitution, declared the AfD a “confirmed right-wing extremist case.”. This legal classification enabled intensified surveillance, including infiltration. This was provisionally suspended in May 2025 by the intelligence agency, which says it will wait for a court ruling on the matter before redetermining its ‘extremist’ nature.

Amid growing concerns, public and political debates over whether to ban the AfD have intensified. The Interior Minister declared the ban as “clear and unequivocal,” adding that the office has “a clear legal mandate to act against extremism and to protect our democracy.” The incoming Interior Minister suggested that it was unlikely that the party would be banned.

The suspension of the classification of the AfD as extremist is provisional, but the possibility of a ban continues to exist. While ban proponents argue that the party represents a real threat to Germany’s liberal democratic order, we want to make a stark warning that banning it could backfire, fueling narratives of martyrdom and persecution and weakening the legitimacy of democratic pluralism.

As Germany revisits this high-stakes question, it must confront the lessons of both its history and its legal obligations under European human rights law. Can democracy protect itself by silencing opposition? Or does such a move endanger the very freedoms it seeks to preserve?

The European Court of Human Rights and the Banning of Political Parties

The ECtHR has been instrumental in shaping Europe’s legal limits on political pluralism. In 1957, the now-defunct European Commission of Human Rights allowed for the prohibition of the German Communist Party. Many of the cases involving political party prohibitions were brought against Turkey. These included parties such as the Turkish Socialist Party and the Communist Party, where the Court showed an unwillingness to dissolve political parties.

There are thus some cases where ECtHR’s message is clear: democracies must be strong enough to tolerate dissenting, even radical, views, so long as those views are expressed and pursued through democratic means. Political parties are not outlawed for what they believe, but for what they do. In Yazar and Others v Turkey (2002), the ECtHR stressed that democratic functionality requires space for political entities to voice proposals that may diverge from government policy or dominant public sentiment. However, this recognition does not translate into the ECtHR’s granting of unlimited or unregulated participation in democracy.

Despite some good outcomes, the Court has also adopted a stringent, militant democratic approach with no substantial assessment of the real harm that may emanate from dissolving political parties. The landmark case Refah of Partisi (the Welfare Party) v. Turkey (2003) remains one of the Court’s most authoritative statements on this issue.

In that case, the Turkish Constitutional Court had dissolved the country’s largest political party, citing its adherence to Sharia as incompatible with the principles of secular democracy. The ECtHR upheld the ban, emphasizing that democracy does not require tolerating political movements that seek its destruction from within. The Court concluded that

“a political party whose leaders incite recourse to violence or propose a policy which does not comply with one or more of the rules of democracy or is aimed at the destruction of democracy and infringement of the rights and freedoms afforded under democracy cannot lay claim to the protection of the Convention against penalties imposed for those reasons.”

The Court emphasized that “the dissolution of Refah served the legitimate aim of preserving secularism which lies at the heart of democratic order in Turkey.” However, as noted in the dissenting opinion by three judges, the party was dissolved solely due to statements and actions by its members, not because of anything in its official program or statute, which, in fact, recognized the fundamental nature of secularism. As the dissent notes:

“There is nothing in its constitution or programme to indicate that Refah was other than democratic or that it was seeking to achieve its objectives by undemocratic means or that those objectives served to undermine or subvert the democratic and pluralistic political system in Turkey.”

Looking at whether a prohibition is proportional to the aim pursued, the Court noted that only 5 of the party’s members temporarily lost their parliamentary seats and leadership roles; the remaining 152 members continued to serve, indicating that the measure, though significant, was not absolute in its effects.

The above decision reflects that the ECtHR has erred on the side of caution, allowing for preemptive measures with a limited substantial assessment of harm while also disregarding the grave backfiring that the restriction of a political party may bring about. As the ECtHR has underlined in another case:

“a State cannot be required to wait, before intervening, until a political party has seized power and begun to take concrete steps to implement a policy incompatible with the standards of the Convention and democracy, even though the danger of that policy for democracy is sufficiently established and imminent.”

What happened after Refah? The New Welfare Party was created in 2018 by the son of the late chairperson of Refah Partisi. In the 2023 general elections, it received only a small number of votes and won just five seats in parliament in a group of 600. Still, the party has been reactivated, and earlier this year, its leader announced he will run in the 2028 presidential elections.

Censorship Didn’t Stop the Nazis. It Helped Them.

While some ECtHR cases have reflected a more cautious stance to the banning of political parties, others embraced a rigid militant democracy approach, without a meaningful evaluation of the actual harm posed by the party (or in some cases, association) in question. Yet history cautions against relying too heavily on repression to safeguard democracy.

Germany, of all countries, is acutely aware of the dangers of political complacency in the face of extremism. But it also knows that censorship and repression can backfire. As history shows, militant democracy can become its own enemy if deployed recklessly. A ban on the AfD would test the limits of democratic self-confidence and risk transforming a political opponent into a martyr.

One of the most persistent myths in political discourse is that the Nazis rose to power because Weimar Germany tolerated too much speech. This is the so-called “Weimar fallacy”: the belief that too much freedom paved the way for fascism. Yet the historical record tells a different story.

During the Weimar Republic, Germany adopted draconian emergency laws to protect democracy against both communists and fascists. However, this did not stop Hitler, but it did drastically affect press freedom. Jacob Mchangama writes that the “constant attempts” to silence Hitler and the Nazis “often helped to increase interest in, and sympathy for, the Nazis.”

When the Nazis took power, they abused the Weimar emergency laws to “strangle the very democracy the laws were supposed to protect” and turned them against their political enemies, trade unionists, communists, Jews, journalists and the clergy. The lesson is stark. Censorship can reinforce the movements it seeks to dismantle, and the tools used to suppress hate can be used, in turn, to enforce it.

Germany has already attempted to ban an extremist party in the modern era. The National Democratic Party (NPD), widely regarded as a neo-Nazi successor organization, was the subject of two attempted bans. The most recent attempt, ruled by the Federal Constitutional Court in 2017, concluded that while the party sought to undermine democratic order, it lacked the capacity to do so. Therefore, a ban was disproportionate.

The NPD later brought its case to the European Court of Human Rights (ECtHR), arguing that it had faced a de facto ban through discrimination and surveillance. While the ECtHR ultimately found the complaint inadmissible, it acknowledged the seriousness of the NPD’s grievances, such as exclusion from public venues, lack of banking services, and suppression of demonstrations, not that it made any difference to the actual result. In short, the NPD was politically ostracized without ever being formally dissolved — a cautionary tale of what happens when democratic systems engage in piecemeal repression without formal justification.

An important example of how restrictive laws play out in practice is the case of Björn Höcke, the controversial head of the AfD in Thuringia. Despite being twice convicted for using Nazi slogans, Björn Höcke's AfD came in first in his state, winning just over 38% of the vote—twice as much as the Christian Democrats.

The AfD, unlike the NPD, is no marginal force. Its potential to reshape the German political landscape is real, and thus the temptation to ban it could be stronger. But the risks are also exponentially higher.

On the other hand, there’s a real possibility that Germany will not opt for an outright ban of the AfD, but instead continue down the more ambiguous, and arguably just as troubling, path of limiting its activities through intensified surveillance and infiltration. Labelling the party as “extremist” reinstates this strategy, one that imposes significant constraints without triggering the legal and political fallout of a formal prohibition.

Banning the AfD might be constitutionally possible, but that does not make it prudent. The party already trades in narratives of elite betrayal, censorship, and persecution. A ban would not silence these claims; it would validate them.

Democracies cannot be defended by imitating the tactics of their enemies. They must show confidence in their values, be that openness, pluralism, and the power of persuasion over prohibition.

Conclusion

While the ECtHR has developed a well-defined framework of proportionality, imminence, and necessity, it frequently treats party dissolution as a manageable disruption rather than a constitutional trauma. There is little evidence in its case law of serious engagement with what banning a major democratic actor actually entails, such as possible cultural rupture, political destabilization, and the legitimization of illiberal instincts within liberal democracies.

With proposals to further marginalize the AfD, Germany faces a volatile situation. This is not the NDP. It commands nearly a fifth of the national vote. An outright ban or an aggressive quasi-ban would not only silence a movement but also destabilize the polity.

History teaches us that censorship in the Weimar Republic did not stop fascism. Instead, as Nadine Strossen has pointed out, its laws fueled Nazi propaganda, casted them as victims, and gave them the tools to silence dissent once in power.

Today, the lesson is the same: rather than fear the democratic process, Europe’s institutions and courts must learn to trust it. The best answer to dangerous ideas is their suppression, exposure, challenge, and ultimate defeat in the public square.

Natalie Alkiviadou is a Senior Research Fellow at The Future of Free Speech. Her research interests lie in the freedom of expression, the far-right, hate speech, hate crime, and non-discrimination.

Justin Hayes is the Director of Communications at The Future of Free Speech and the Managing Editor of The Bedrock Principle.

Why do some “offensive” posts remain online while others get removed? Meta’s Oversight Board, a quasi-Supreme Court for Meta platforms like Facebook and Instagram, has just handed down a batch of rulings that might raise more questions than answers. These decisions show how well-intentioned rules can punish the very voices they aim to protect and feed the societal tensions they hope to calm.

On April 23, the Oversight Board (OB) published its most recent decisions following Meta’s policy changes. Each ruling provides the public with insights into how some of the world’s biggest platforms enforce their speech policies. These decisions are always welcome, particularly when it comes to the broader status quo where social media platforms routinely remove ‘lawful but awful’ speech.

At the same time, the developments point to an inconsistency in treatment amongst different types of hate speech, raising significant concerns about equal protection and the broader implications for social cohesion. Using the principles of International Human Rights Law (IHRL), empirical data, and contemporary legal debates, this piece will examine the real-world consequences of these decisions and their implications for the future of free speech online.

Here’s the bottom line: establishing a high bar for limiting speech is an absolute necessity, not just a welcome recommendation. That high threshold needs to apply to all types of hate speech equally, be that racist speech or transphobic speech, for example. Without this framework, we risk stifling important public debate, discounting already underrepresented voices, and exacerbating an already polarized society.

The OB Sides with Free Speech in Cases Involving Gender Identity and the Apartheid Flag

In 2024, two Meta users misgendered transgender individuals in posts about bathroom use and female sports. These posts were reported for hate speech, but both were left up by Meta. Users appealed, prompting review by the OB, which supported Meta’s decision. Although the content involved misgendering transgender individuals, the majority of the OB found that, while the content in question could be considered offensive, it did not incite violence or discrimination, and the issue at stake was a matter of public concern.

In agreeing with Meta’s position on the case, the OB cited parts of the public comment put forth by The Future of Free Speech (FoFS), in which we urged the OB to adhere to Meta’s decision of non-removal. This position emanates from IHRL, specifically Article 20(2) of the International Covenant on Civil and Political Rights (ICCPR), which prohibits “any advocacy of national, racial or religious hatred that constitutes incitement to discrimination, hostility or violence'“ as well as the six-part threshold test set out by the Rabat Plan of Action (RPA).

In addition to law, we emphasized the importance of protecting open public debate on controversial issues, such as gender identity, warned against the democratic risks of censorship, and recommended consistency across Meta’s platforms. Suppressing public debate on issues like gender identity, even when it includes offensive language, risks chilling free expression on one of the most contested issues of our time. Significantly, this chill affects not only critics of prevailing norms but also transgender individuals themselves, who may feel less able to participate in discussion if the space becomes over-regulated or adversarial.

The OB decision referred to a case in May 2024, where a user posted a video on Instagram of themselves performing in a drag show, wearing a red, glittery outfit. The caption thanked collaborators and described themselves as a “faggy martyr.” Meta removed this post on hate speech grounds and reinstated it after the OB brought the case to Meta’s attention. This illustrative case highlights the challenges that marginalized groups face when it comes to content moderation. In response to stricter legislative demands and the threat of substantial fines, platforms often adopt a “better safe than sorry” stance, leading to overly stringent content moderation.

Furthermore, the use of AI may result in the biased enforcement of companies’ content policies. Whether it’s from a lack of data or biased training datasets, members of minority communities can potentially be silenced when AI platforms are unable to detect the nuance of language or linguistic reclamation. For the LGBTQ+ community, terms such as “dyke”, “fag” and “tranny” are a way of reclaiming power and a means for preparing members of this community to “cope with hostility.”

But this case shines a light on a deeper problem: the algorithms and policies built to spot “hate speech” often muzzle people who are simply speaking in their own voice. The trans community isn’t the only casualty. Tweets written in African American English are flagged as “offensive” almost twice as often as other tweets—clear evidence that bias still lurks in the code. If social platforms want to be the modern town square, they must guarantee that the very groups historically pushed to the margins can speak as freely as anyone else.

Similarly, in two cases involving images of South Africa’s apartheid-era flag, the OB upheld Meta’s decision to keep the posts online. While recognizing the flag’s painful legacy, the Board concluded that removing the content was not the least restrictive means available and affirmed that, even if offensive, the posts remained protected under international human rights law (IHRL). Interestingly, and reflective of the strong adherence to free speech that marked this round of cases, the OB’s majority decision to keep the posts up was made despite the finding that they breached Meta’s rules on hateful ideologies. It emphasized that this was done to remain in line with IHRL and called for greater clarity in how this Community Standard is defined and applied, reflecting a conflict between those rules and IHRL.

The OB’s Inconsistent Application When It Comes to Anti-Migrant Speech

In two cases involving anti-migrant posts from Poland and Germany, the OB overturned Meta’s original decision and ordered the content removed. The posts included a racist slur and sweeping claims about migrants as sexual predators. The majority concluded that, given the charged political climate ahead of key elections, the content posed a heightened risk of inciting discrimination and violence.

While the OB refers to the Rabat Plan of Action (RPA) and its six-part threshold test regarding the context of the post (in this case, the European elections), it did not conduct an analysis of the content in question under each of these points. While potentially unpleasant and offensive, it is unclear how this speech meets the threshold required for removal under Article 20(2) elucidated by the RPA.

Let’s take a look at each of the six RPA factors:

• Statement Context: While election periods are often marked by polarizing immigration debate, and while this may raise sensitivity, not all heightened political tension justifies speech restriction.

• Speaker’s Position or Status: While the status of the speakers (a political party and a public page) extends the message's reach, their institutional weight does not reach the level of state actors or influential figures, who could incite direct action.

• Intent to Incite Audience against A Target Group: While inflammatory, the posts appear to constitute populist rhetoric rather than a direct call to action for discrimination or violence against particular groups.

• Content and Form of the Statement: Although offensive and stereotypical, the content and form do not clearly advocate violence or hatred that satisfies the legal definition of incitement.

• Extent of Dissemination: The extent of the speech act was not demonstrably wide, and there was no virality or mass dissemination associated with the content.

• Likelihood of Harm: The lack of dissemination would lower the risk that harm would result from the speech.

While it’s beyond the scope of this essay to perform a complete analysis, it is clear that these factors were not fully considered in the OB’s judgment. Its decision noted the ‘unpredictable nature of online virality’ as a justification for Meta taking a more cautious approach to moderation. Notably, former Special Rapporteur on the Freedom of Opinion and Expression David Kaye highlighted that “international human rights standards can guide such policies, while the virality of hateful content in such contexts may require rapid reaction and early warning to protect fundamental rights.”

Nevertheless, this does not mean that the essence of the RPA must be diluted. More importantly, even if the OB believes these posts meet such tests, it would have been crucial for it to explain why. Pre-emptive speech restriction without adequate proof of imminent harm is not in line with IHRL, the standard by which the OB seeks to abide.

Elections, particularly in democratic states of the European Union, can by no means serve as a carte blanche for restricting offensive speech. In fact, Eric Heinze argues that in long-standing and prosperous democracies, characterized by sound institutions, “despite decades of pro-ban law and policy… no empirical evidence has, in any statistically standard way, traced hatred expressed within general public discourse to specifically harmful effects.”

The Problems with Restricting Hate Speech

Based on IHRL and other legal standards for determining the threshold of hate speech, the OB got two out of three decisions correct. While the gender identity and apartheid-flag cases reflected positive developments, the drop in speech thresholds when it came to anti-migrant speech does not show a consistent position by OB.

These cases reflect the legal, philosophical, and practical issues involved in regulating hate speech. Law Professor Eugene Volokh has coined the term “censorship envy,” referring to the resentment that arises when one group is shielded from criticism while others are not.

On a very basic level, the above cases could be interpreted as a protection of transphobic and white supremacist speech on the one hand but as a repression of anti-immigrant speech on the other. This inconsistent application can precipitate violence or social unrest, particularly when individuals perceive their expressive freedoms as being unduly restricted while other speech receives more protection. In such cases, some may resort to more confrontational or even violent forms of protest.

Within this context, protecting free speech, even when it is offensive or unsettling, can serve as a critical safety valve, potentially mitigating the risk of more dangerous outcomes. A key strategy in addressing far-right violence lies in reducing societal polarization, which implies that “openness and dialogue might work better than public repression, stigmatization or aggressive confrontation.” Research has shown that the rise of far-right extremism in Western Europe is rooted in a confluence of factors, including high levels of immigration, limited electoral success of radical right-wing parties, and the “extensive public repression of radical right actors and opinions.”

Moreover, restricting speech considered hateful, without meeting the strict tests of necessity and proportionality set out by IHRL, can lead to counterproductive results, such as pushing hateful attitudes underground. This allows extremists to adopt martyrdom narratives and contributes to the polarization of societies and social conflict through negative repercussions and backlash effects. While the OB’s citation of academic research was promising, its future decisions could benefit from the further use of such work, including empirical findings that reflect the actuality of situations.

Rather than embracing blunt instruments of censorship, content moderation should be based on IHRL and the RPA, provided that context does not, as was the case with anti-migrant speech, constitute a carte blanche for removal. Meta and the OB should, in detail, apply tests and thresholds to reach necessary, proportional, and, therefore, legitimate conclusions. Ensuring free speech on platforms in instances where the RPA conditions are not met protects the vital democratic function of speech since it allows for challenging, questioning, and dissenting opinions, even when they are offensive.

When hate speech poses (non-violent) challenges, especially for marginalized communities, censorship is not the answer. Censorship risks eroding democratic trust, stifling minority voices, and replacing open dialogue with enforced conformity. There is nothing stable, sustainable, or long-term in this approach. Quite the opposite. A free society cannot exist where controversial ideas are automatically equated with harm or, even worse, equated with harm in some types of hate speech (anti-migrant) and not others (transphobia and white supremacy).

In a free society, harm must be demonstrated, not assumed, and the default remedy for bad speech should be more speech, not silence. Counterspeech and counternarratives are essential tools for people to address harmful speech without censorship. Studies show that community-driven responses to hate speech, such as satire, rebuttals, or solidarity campaigns, are more effective and less polarizing than bans.

Above all, any restriction must be necessary and proportionate. As the former UN Special Rapporteur on Freedom of Expression emphasized, removing content should always be the least intrusive measure available, especially when dealing with controversial but lawful speech. Meta and the OB must take great care in how restrictions can backfire, leading to dismal consequences on individual, group, and societal levels.

Natalie Alkiviadou is a Senior Research Fellow at The Future of Free Speech. Her research interests lie in the freedom of expression, the far-right, hate speech, hate crime, and non-discrimination.

Share

The European Union’s sweeping online safety law continues to make headlines — some controversial — as we learn more about how it will be implemented and enforced in practice. That law, the Digital Services Act (DSA), has been hailed as a novel “road map” for reigning in the “Wild West” of the internet that has allegedly been overrun with “torrents” of hate speech and illegal content.

But how exactly will these new regulations ensure that platforms deal with illegal hate speech without stifling freedom of expression online? Recent developments raise some significant red flags.

In 2016, the European Commission introduced the Code of Conduct on Countering Illegal Hate Speech Online, a voluntary agreement with major tech companies aimed at ensuring the swift removal of illegal online hate speech. It marked an early attempt at co-regulation in the absence of binding EU-wide legislation.

Over time and with the adoption of the DSA, this initiative evolved into a more formalized and enforceable framework, culminating in the Code of Conduct on Countering Illegal Hate Speech Online + (Code+). In January 2025, this code was integrated into the Digital Services Act (DSA). The signatories to date include Facebook, Instagram, X, YouTube, TikTok, Dailymotion, Jeuxvideo.com, LinkedIn, Microsoft-hosted consumer services, Snapchat, Rakuten Viber, and Twitch.

The first issue of concern is the extent to which the Code+ will, in practice, be voluntary. The DSA explicitly states that the European Commission and the European Board for Digital Services shall “encourage and facilitate the drawing up of voluntary codes of conduct at Union level to contribute to the proper application of this Regulation.” Recital 103 of the DSA notes while codes should be measurable and subject to public oversight “this should not impair the voluntary nature of such codes.”

Nevertheless, Article 45 of the DSA grants the Commission and the European Board for Digital Services oversight powers, allowing them to exert pressure on signatories for Code compliance. Specifically, Part 4 of this article stipulates that “in the case of systematic failure to comply with the codes of conduct, the Commission and the Board may invite the signatories . . . to take the necessary action.”

What exactly constitutes a “necessary action” is never defined, but the infrastructure of the Code+ within the ambit of the DSA could effectively transform ‘voluntary commitments’ into more rigid obligations under regulatory scrutiny. This framework could end up creating a de facto enforcement mechanism, blurring the line between voluntary adherence and regulatory compulsion. Moreover, the DSA requires platforms to mitigate “systemic risks” while the Code+ refers to Article 35 (1) (h) of the DSA, which renders adherence to voluntary codes as an appropriate risk mitigation measure in the framework of Very Large Online Platforms and Very Large Online Search Engines.

In light of the above nuances, and while the Code+ is framed as a voluntary instrument creating “voluntary commitments,” it could be argued that, in practice, its implementation will not pan out to be completely voluntary. While time will tell, particularly in terms of the rigor and findings of the monitoring reports, civil society organizations and other stakeholders must remain vigilant to ensure that the voluntary nature of such codes is retained.

Additionally, several issues that plagued the 2016 Code persist, including the absence of a clear definition of hate speech, the risk of excessive content removal, and the reliance on trusted flaggers. These challenges could become even more problematic with the integration of the Code+ into the DSA framework.

A Lacking Definition of Hate Speech

A key issue with the DSA is its lack of a clear definition of hate speech. The DSA obligates digital platforms to assess and remove illegal content, including hate speech and unlawful discriminatory content, but does not specify what constitutes “illegal content.” Instead, it leaves platforms with considerable discretion, requiring them to rely on national laws aligned with EU regulations or interpret EU law directly.

The Code+ follows the precedent set by the 2016 Code of Conduct, using the Framework Decision on Racism and Xenophobia and national laws, transposing it as a definitional benchmark. However, this reliance presents several challenges. The Framework Decision is a criminal law instrument that sets a high threshold for prohibition, focusing on incitement to violence or hatred. It also does not explicitly define hate speech and only protects characteristics such as race, color, religion, descent, and national or ethnic origin—offering a narrower scope than most social media policies.

Despite the reference to the Framework Decision and national laws, platforms must remove content that violates their policies as well as national laws more broadly and not only those transposing the Framework Decision. However, broad platform policies that extend, amongst many others, to expressions of contempt and stereotypes are often a far cry from international human rights standards. Also, variations in national legal frameworks create uncertainty for platforms operating across Europe.

Given the complexity and contentious nature of hate speech regulation, both the Code+ and the DSA would have benefited from a precise definition. Nevertheless, even with such a definition, tasking private companies that are not bound by International Human Rights Law to ascertain the limits of the fundamental freedom of expression is a serious problem in the current infrastructure.

More Over-Removal?

A key concern arising from the Code+ and its role in the DSA’s ambit is the over-removal of content. It is essential to first acknowledge that the DSA mandates platforms to remove illegal hate speech “without undue delay.” However, the Code+, like its predecessor, imposes a 24-hour deadline for content removal. In terms of quantitative requirements, the signatories commit to review at least 50% of notices received. They will “apply their best efforts to go beyond this target” and aim at least 67% (two-thirds) of the notices.

In a 2021 report, the UN Special Rapporteur on the Promotion and Protection of the Right to Freedom of Opinion and Expression, Irene Khan, cautioned that increased platform liability may incentivize intermediaries to engage in the excessive removal of content. As she poignantly noted, “By compelling social media platforms to police speech, they create a risk that companies will zealously over-remove material and undermine free speech.”

To deal with the sheer amount of online content but also due to enhanced obligations, platforms rely on automated content moderation. As emphasized in a Council of Europe report, automated mechanisms have a direct impact on freedom of expression, raising concerns regarding the rule of law, particularly in terms of legality, legitimacy, and proportionality. The Council of Europe cautioned that the increased reliance on AI for content moderation could lead to over-blocking, thereby endangering freedom of expression. Additionally, in a 2020 study, Gorwa, Binns, and Katzenbach argue that the growing use of AI risks exacerbating the existing opacity of content moderation, complicating online justice, and “re-obscur[ing] the fundamentally political nature of speech decisions being executed at scale.”

In other words, automated mechanisms inherently lack the capacity to grasp the nuance and context of language and human communication. As my colleague Jacob Mchangama pointed out in a previous article, both human and automated content moderation led to the removal of several satirical and irony-filled posts on Facebook and Instagram due to a failure to understand such context.

With the combined pressure to remove content at such a large scale and the potential legal repercussions of failing to do so, it is safe to assume that platforms will inevitably over-remove content. In a 2024 report, The Future of Free Speech found that a significant proportion of content removed by Facebook and YouTube in France, Germany, and Sweden was, in fact, legally permissible. Depending on the dataset, between 87.5% and 99.7% of deleted comments fell within the bounds of legality.

Monitoring Reporters and Trusted Flaggers

The results of the monitoring exercises raise another issue of concern pertaining to “trusted flaggers” and “monitoring reporters.” Both the DSA and the Code + refer to trusted flaggers, which are entities appointed by national authorities and have expertise in detecting, identifying, and notifying illegal content.

The Code + also provides for a network of “monitoring reporters,” which are “not-for-profit or public entities with expertise on illegal hate speech in at least one EU Member State and approved by the Commission and the Signatories to participate in the Monitoring Exercise.” The monitoring reporters will be conducting annual monitoring exercises to ascertain the extent to which the signatories are complying with the Code+.

Monitoring reporters may also be trusted flaggers. Under the original Code, trusted flaggers were primarily equality and non-discrimination bodies, along with some national institutions. Notably, there was an absence of free speech organizations among the trusted flaggers, as can be seen in the lists of flaggers in each of the seven monitoring exercises.

This composition of groups could lead to a bias toward content removal since it creates an imbalanced oversight mechanism that risks undermining fundamental rights. The absence of a free-expression perspective means that decisions about content removal may disproportionately prioritize offense avoidance over legitimate discourse, satire, or political critique, for example. Moreover, free speech organizations play a critical role in scrutinizing whether removals align with principles of legality, necessity, and proportionality, ensuring that restrictions on expression do not erode democratic debate.

It remains to be seen which organizations (at least one per country) will undertake the monitoring exercises, but a better approach would be to adopt a more holistic composition of organizations, especially compared to the 2016 Code. If both non-discrimination and free speech advocates participate, it will create a more balanced, transparent, and rights-respecting content moderation framework. This recommendation would not solve the inherent issues and controversies of the EU’s approach to content moderation, but it could certainly improve the status quo.

The fact that monitoring reporters can also be trusted flaggers appointed under the DSA creates additional issues. Under Article 22 of the DSA, trusted flaggers are appointed by national authorities, which raises concerns about potential governmental overreach, especially in countries with authoritarian tendencies.

For instance, in countries where the government has been criticized for undermining democratic institutions, their appointment of trusted flaggers could lead to the suppression of dissenting voices under the guise of combating hate speech. Moreover, the opacity surrounding how national regulators appoint trusted flaggers makes it challenging to assess whether these actors are impartial and adequately representative.

Counter-Narratives

The Code+ also refers to enhanced cooperation between stakeholders to raise awareness of illegal hate speech, including how to report it and prevent its spread. From a free speech perspective, the inclusion of counter-narratives in the Code+ is particularly significant.

The 2016 Code acknowledged the role of civil society organizations in countering online hatred through the development of alternative narratives and pledged collaboration between signatories and the European Commission to promote independent counter-narratives. However, in practice, the emphasis remained mainly on monitoring and content removal, leaving the implementation of this commitment unclear. With the Code+, the counter-narrative approach has been reaffirmed, and ideally, this time, greater resources and attention will be dedicated to developing meaningful tools and initiatives in this area, which have proven effective for combatting hate speech (online and offline).

Free Expression Needs Consideration

Nowhere does the Code + refer to freedom of expression. This critical absence reflects the deterioration of this fundamental freedom under the current European approach to the moderation of online content.

The Code+’s addition of a timeframe creates additional pressure on platforms to remove content, while the continuation of the monitoring exercise increases the chances that considerations for free expression will be wholly absent from how the code works in practice.

Let’s hope that free expression advocates and experts will have a seat at the table. If not, we should continue to demand that this fundamental right be weighed equally with the other objectives of the DSA as it comes into full focus.

Natalie Alkiviadou is a Senior Research Fellow at The Future of Free Speech. Her research interests lie in the freedom of expression, the far-right, hate speech, hate crime, and non-discrimination.

Share The Bedrock Principle