2. Fighting with mis/disinformation

In this section, I will briefly discuss the existing methods of combating misinformation and disinformation and situate my proposal within that broader framework. At a general level, such measures can be divided into two groups: those that address false information after it has already spread, and those that aim to prevent its dissemination before it occurs. The first group focuses on what to do once misinformation is already in circulation – how to identify, debunk, or counteract it. The second group seeks to prevent misinformation from being published in the first place, or to mitigate its potential destructive impact before it reaches an audience (after all, information ceases to be harmful if no one believes it). Both kinds of measures are necessary. In this paper, however, I defend the thesis that there should exist incentive structures – grounded in the criminal law apparatus – that discourage the publication of disinformation. This proposal is not intended to stand alone but rather to supplement existing approaches.

Numerous measures have been proposed to combat misinformation and disinformation after its publication. Those ideas focus on platforms that vary significantly, from deleting information and removing accounts that spread it (Bursztynsky Reference Bursztynsky2021). The content is, after intervention, not available to other users, and the author of the intervention cannot post again on that platform. Fighting with mis/disinformation does not need to involve the deletion of the content. For example, flagging problematic content as false might lessen the influence of that information. (see, e.g., Kim et al. Reference Kim, Tabibian, Oh, Schölkopf and Gomez-Rodriguez2018; Aruguete et al. Reference Aruguete, Bachmann, Calvo, Valenzuela and Ventura2025; Steensen et al. Reference Steensen, Kalsnes and Westlund2024). A similar strategy could involve promoting, next to the false information, other sources with a credible explanation of the phenomenon (see, e.g., Alemanno Reference Alemanno2018). The other mechanism, the reduction of the visibility of problematic content, might also be significantly limited (Cotter et al. Reference Cotter, DeCook and Kanthawala2022).

Education is a measure that belongs to both categories. Implemented before the spread of misinformation, it can prevent people from publishing false information (“I will not post this because I know it is untrue”) or help neutralize potentially harmful content (“I do not believe this information because I know it is false”). Educational efforts can also occur after the publication of misinformation (“I no longer believe this information because I have been persuaded that it is false”). However, educational strategies face significant practical challenges, which justify considering additional preventive measures. Before turning to these, let us first examine where the core problem lies.

One might think that education and the explanation of why certain information is false could help; however, this is not always the case. People are not easily persuaded (Gottlieb Reference Gottlieb2016; Klein Reference Klein2021), and meta-analyses suggest that attempts at debunking misinformation are often unsuccessful (Chan and Albarracín Reference Chan and Albarracín2023). Studies show, for instance, that even a few minutes of exposure to false information about vaccine safety can decrease people’s willingness to vaccinate (Betsch et al. Reference Betsch, Renkewitz, Betsch and Ulshöfer2010). Moreover, the process of debunking false information can, in some cases, lead to undesirable outcomes: more people may become exposed to harmful ideas, thereby increasing the visibility of information that should not be epistemically available (Mosleh et al. Reference Mosleh, Martel, Eckles and Rand2021).

Another issue concerns the motivations behind spreading misinformation and disinformation, for example, religious beliefs, political preferences, fame, money, and even geopolitics (Dhama et al. Reference Dhama, Sharun, Tiwari, Dhawan, Emran, Rabaan and Alhumaid2021; Ahmed Reference Ahmed2021; Scott Reference Scott2020; Size Reference Size2020; Melchior and Oliveira Reference Melchior and Oliveira2024). If someone already knows that the information is false (or at least recognized as false by the mainstream), educational efforts are likely to be ineffective. Moreover, as mentioned above, the public process of debunking such information may unintentionally expose others to harmful content, thereby amplifying rather than reducing its impact.

What lesson can be drawn from this? It may be better not to be exposed to false information at all, because once it becomes available, its influence cannot simply be undone. This suggests that certain kinds of false information should perhaps not be available at all. Dennis and Lindberg, in their paper about the threat of misinformation to public health, argue that there is “an urgent need for primary prevention, […] Anything less means misinformation – and its societal consequences – will continue to spread” (Denniss and Lindberg Reference Denniss and Lindberg2025: 1) The core proposition of this paper is coherent with that way of thinking and aim to limit the spread of misinformation by creating incentives not to publish it in the first place.

One might ask why I classify criminal responsibility as a preventive measure. After all, criminal responsibility is typically imposed after a crime has been committed. However, I refer here to criminal law as a preventive tool in a utilitarian sense. In brief, the justification for criminal law can be either backward-looking – we punish because offenders deserve it – or forward-looking – we punish to prevent future crimes (on theories of punishment, Canton Reference Canton, Focquaert, Shaw and Waller2020). Both aspects are present in most legal systems, but in this context, I understand the criminalization of the abuse of epistemic power to spread misinformation primarily as a preventive measure. It is intended to serve as a deterrent for those who hold epistemic power and might otherwise be inclined to publish or engage with potentially harmful content (on deterrence, see Apel and Nagin Reference Apel, Nagin and Tonry2011; Robinson and Darley Reference Robinson and Darley2004). This deterrent effect can operate regardless of the individual’s initial motivation: whether the person believes the information to be true or not, the mere possibility of criminal liability may function as a negative incentive to refrain from posting it. Of course, if such an offense were introduced, some individuals would ultimately be punished under it; yet these cases would also serve as a lesson to others, reinforcing the broader preventive purpose. In practice, criminal responsibility applied to a few can have a much wider social effect.

3. An objective approach to the criminalization of disinformation

Using criminal law to combat false information predates the problem of fake news spread over the Internet. One kind of law that could be interpreted in that way is the crime of Holocaust denial, which is part of the legal system in various countries (see, e.g., Kahn Reference Kahn2004; Teachout Reference Teachout2005). In brief, this crime might be committed when someone claims that the Holocaust did not happen. It could be considered, then, as anti-fake news regulation, because it protects the truth about a particular historical event, and if someone publicly claims differently, they could be criminally charged. For the record, there are voices that such a law should be repealed (see, e.g., Singer Reference Singer2016).

Nevertheless, the idea of introducing criminal responsibility to fight against false information is not new, and the usual formulation of the problem is to focus on spreading concrete categories of fake news, for example, disinformation in the electoral period or publishing deep fake porn (see, e.g., Beech Reference Beech2018; Statt Reference Statt2019; Lecher Reference Lecher2019; Gold and Washington Reference Gold2025).

I will illustrate the core idea of this paper using a public health example that I have developed in my other work. These issues are, to some extent, independent. One might agree that there is a need to introduce criminal responsibility for superspreaders with epistemic power without necessarily endorsing my specific formulation of the crime of spreading medical disinformation. However, for the purposes of this discussion, it will be easier to present the concept of abuse of epistemic power for the spread of disinformation with a concrete example in mind.

Elsewhere, I proposed the criminalization of spreading medical fake news (Mamak Reference Mamak2021). The core idea stemmed from the observation that the Internet has become a powerful medium for disseminating false health-related information, which has led to the rise of vaccine hesitancy. In that article, I focused particularly on anti-vaccination claims, such as the false allegation that childhood vaccines cause autism. Empirical data at the time showed an increase in the incidence of diseases that could have been prevented through vaccination, suggesting that misinformation had tangible, harmful public-health effects.

In that context, I examined the limitations of non-criminal strategies for countering fake news – such as education, fact-checking initiatives, and self-regulation by digital platforms – and argued that if these measures prove ineffective, a narrowly tailored criminal provision may be justified. The proposed offense targeted the public dissemination of information that is evidently discrepant with established medical knowledge (“Whoever publicly disseminates information evidently discrepant with medical knowledge is subject to a penalty.”), aiming to deter the circulation of health-related falsehoods that endanger others. For the record, it should also be noted that not all false information about medicine would qualify as prohibited; only that which has the potential to cause substantial harm should fall within the scope of the restriction (Mamak Reference Mamak, Faintuch and Faintuch2022).

I acknowledged that such a measure could be perceived as a restriction of freedom of expression. However, as I argued there, this type of restriction can be constitutionally defensible, since constitutional systems typically permit limitations on certain freedoms when they conflict with other fundamental goods – here, the protection of public health and human life. The proposed approach thus sought to balance freedom of speech with the state’s duty to safeguard citizens against serious health risks created by the deliberate or reckless spread of medical misinformation.

The proposal I will now discuss builds on ideas I first formulated before the recent pandemic. Observing the spread of COVID-19 and the accompanying wave of misinformation convinced me that pandemics require special treatment. Future outbreaks are likely, and one of the key factors influencing the effectiveness of any response will be the quality of information available to the public. Even if safe and effective vaccines are developed, they may not be widely used if misinformation portrays them as dangerous or deadly.

In my recent publication, I revisited my earlier proposal for the criminalization of media-based fake news and adapted it to the specific context of pandemics (Mamak Reference Mamak, Faintuch and Faintuch2025). I proposed the following offense:

The difference between this provision and the earlier, more general one is twofold. First, it applies a lower standard of certainty regarding what counts as medical knowledge. Second, it is limited to periods of pandemics. During such times, controlling the spread of information is crucial, as it can directly affect public health outcomes. The COVID-19 pandemic revealed how fragile public trust in vaccines can be, and widespread disinformation may lead to preventable deaths. Therefore, restricting the circulation of vaccine-related information during pandemics may be justified.

Importantly, the lower threshold of certainty about medical knowledge does not mean that any publication referring to medical matters qualifies as “medical knowledge.” Rather, this term should be understood as information recognized as such by experts. This assessment may depend on factors such as the place of publication (e.g., reputable versus predatory journals), the methodology employed, and the general acceptance of findings within the scientific community.

In sum, it is useful to ground future deliberations on the abuse of epistemic power in specific examples of misinformation. There are many possible cases to consider; however, in this paper, I have focused on two offenses related to public health issues.

4. A subjective approach to criminalization of disinformation: epistemic power

We now turn to the central part of this paper, where I present the justification for a subjective differentiation of criminal liability for abuse of epistemic power. The argument unfolds in three steps. First, I will show that legal systems sometimes differentiate criminal responsibility based on the subjective characteristics of the individual. I will use the example of public officials, who can be punished for abuse of power – conduct that constitutes a crime only when committed by those holding power, not by ordinary citizens. Second, I will argue that epistemic power is a meaningful category and that we can distinguish online actors according to whether or not they possess such power. Third, I will demonstrate that the law already differentiates legal evaluation according to the scale of the actor’s influence. Smaller actors are often treated differently from larger ones; this is evident, for instance, in the regulation of abuse of market power. Together, these three steps aim to show that criminalizing only those who abuse epistemic power could be a defensible strategy for combating disinformation. The points presented address a potential counterargument that the proposal might be considered unacceptable because it violates the principle of equality before the law. This section explains why it could nonetheless be acceptable – even in the context of criminalization – to treat different actors differently when they engage in the same act of disseminating disinformation.

5. Crime of abuse of epistemic power for spreading disinformation

In this section, I outline how the idea of criminalizing the abuse of epistemic power could be conceptualized and operationalized. I present and explain a legislative proposal that serves as a basis for a more in-depth exploration of the underlying idea. This proposal should not be understood as a fully developed legal solution ready for implementation, but rather as a starting point for further discussion and refinement.

The intention is not to suggest that every instance of disinformation disseminated by individuals with epistemic power should be criminalized. Instead, the aim is to consider whether, in cases where criminalization of specific categories of false information is already justified, such regulation might be limited to those who possess epistemic power. In other words, before addressing who should fall within the scope of criminal liability, we must first establish a substantive justification for prohibiting particular categories of disinformation.

Let us assume, for the sake of discussion, that there is broad agreement that a specific category of false information – for instance, disinformation posing a threat to public health – is sufficiently dangerous to warrant criminal regulation. However, it is also likely that a broad or indiscriminate criminalization of such content would be socially and politically unacceptable, as it could give rise to concerns about widespread censorship. (For present purposes, I set aside the question of whether such concerns are justified.)

In such a scenario, a pragmatic way forward would be to narrow the subjective scope of regulation – that is, to limit who can be held criminally liable under the proposed framework. This could be achieved by confining the regulation to individuals who possess epistemic power, thereby targeting those whose speech has the greatest potential impact on public understanding and behavior. The proposed limitation could be formulated as follows:

I need to unpack this little bit. The concept of epistemic power remains ambiguous and requires clarification within the context of regulation. As previously noted, epistemic authority is multidimensional; follower counts alone are insufficient and should be complemented by indicators such as average reach, engagement rate, or the individual’s social role (e.g., journalist or medical professional) (Bartsch et al. Reference Bartsch, Neuberger, Stark, Karnowski, Maurer, Pentzold, Quandt, Quiring and Schemer2025). Nevertheless, for regulatory purposes, epistemic power must be operationalized in a manner that is easily understandable, particularly for individuals who might face criminal liability.

The most straightforward way to translate the concept of epistemic power into a clear legislative standard is to express it numerically. For instance, Conde and Casais (Reference Conde and Casais2023) classify Instagram influencers into categories based on follower count: micro-influencers (1000–100,000 followers), macro-influencers (100,000–1 million), and mega-influencers (over 1 million). Other researchers adopt stricter definitions; for example, Park et al. (Reference Park, Lee, Xiong, Septianto and Seo2021) define micro-influencers as those with at least 10,000 followers. Interestingly, their study also finds that smaller influencers are often more persuasive than mega-influencers, as they tend to be perceived as more authentic. Similar findings appear in Kay et al. (Reference Kay, Mulcahy and Parkinson2020), which suggests that smaller influencers may exert greater persuasive power than larger ones, at least within marketing contexts.

Walter et al. (Reference Walter, Föhl and Zagermann2025) further compare micro- and mega-influencers across dimensions such as trustworthiness, expertise, attractiveness, authenticity, and similarity, showing that micro-influencers (over 10,000 followers) are generally perceived as more authentic and trustworthy, and outperform mega-influencers particularly in perceived similarity.

Taking these findings together, it seems unwise to set the regulatory threshold for epistemic power too high, given that micro-influencers can possess substantial persuasive influence, sometimes exceeding that of mega-influencers. I therefore propose a threshold of 10,000 followers as a preliminary benchmark. This figure should, of course, be context-sensitive, as follower counts may vary in meaning across platforms, countries, and audiences. Nonetheless, a single clear numerical threshold provides clarity and legal certainty, allowing for an objective distinction between those who possess significant epistemic power and those who do not.

While there may be sound reasons to adjust this threshold, the proposal offered here should be viewed as an invitation to discussion – a starting point for developing a workable and transparent standard for regulating epistemic power.

Focusing solely on follower count reduces epistemic power in the proposal to social media presence. Yet, epistemic power is a much broader concept, traditionally encompassing individuals and groups who shape collective understanding and belief beyond digital platforms. These include, for example, scientists, religious leaders, and public intellectuals – figures who may hold considerable epistemic influence even in the absence of a substantial online following (on epistemic authority, see Zagzebski Reference Zagzebski2012).

There is, of course, potential overlap between these categories. A scientist or religious leader might also maintain an active online presence, thereby combining traditional and digital forms of epistemic authority. Nevertheless, this is not always the case; many individuals wielding significant epistemic influence do so primarily through offline institutions, such as academia, religious organizations, or the media.

The starting point of the current proposal is social media, as they represent the primary channels through which misinformation and disinformation spread today (e.g., Denniss and Lindberg Reference Denniss and Lindberg2025). However, future regulatory frameworks could consider how to incorporate other sources of epistemic authority that operate outside or alongside digital networks. Expanding the scope in this way would allow for a more comprehensive understanding of epistemic power as a multi-domain phenomenon, ensuring that regulation keeps pace with the evolving landscape of knowledge production and dissemination.

A further issue concerns the timing of the follower count used to determine epistemic power. The number of followers must be assessed at the moment when the act is committed or the content is published. It is conceivable that a particular post – perhaps one containing misleading or sensational information – could itself lead to a rapid increase in followers, pushing an account above the 10,000 threshold. However, such a change in status should not retroactively affect responsibility for that earlier act.

In other words, if an individual had fewer than 10,000 followers at the time of posting, they would not fall within the scope of the proposed regulation for that specific content, even if their follower count subsequently exceeded the threshold as a result of the post’s popularity. Only content published after the account surpasses the threshold would fall under the new regulatory regime.

This temporal clarification is essential for ensuring legal certainty and avoiding retroactive liability, which would be inconsistent with fundamental principles of criminal law. A clear rule – tying epistemic power to the number of followers at the time of the act – allows individuals to foresee when their speech becomes subject to stricter regulatory standards. It also prevents arbitrary enforcement and ensures that the regulation targets those who, at the time of communicating information, already hold a significant and stable epistemic influence over others.

6. Free speech issues

Now, I want to spend some time on freedom of speech, which is highly connected with regulations on fake news (cf. Nuñez Reference Nuñez2020; Helm and Nasu Reference Helm and Nasu2021; Jacobs Reference Jacobs2022). Helm and Nasu argue, however, that “[…] that some level of restriction on freedom of expression is inevitable due to the need to discourage the creation and distribution of fake news, rather than just preventing its spread” (Helm, and Nasu, Reference Helm and Nasu2021: 303).

I mentioned that this proposal might be more acceptable from the perspective of freedom of speech than other propositions that call for the use of criminal law to mitigate the spread of disinformation.

Now, I want to explain why. Limiting criminal liability for disinformation to speakers who wield epistemic power is less invasive of freedom of expression than a blanket offense that applies to everyone, because it is narrowly tailored to the locus of greatest potential harm while preserving the expressive space of ordinary users. First, it advances the legitimate aim (e.g., protecting public health) by focusing on actors with amplified reach and persuasive capacity: the law targets the few whose speech predictably causes outsized effects. Second, it reduces chilling effects on everyday participation in public debate; most citizens can speak, criticize, and err without fearing criminal liability, which safeguards the “breathing space” essential to democratic discourse and scientific contestation. Third, using a clear, ex ante threshold (e.g., 10,000 followers, assessed at the time of the act) provides foreseeability and legal certainty, enabling affected speakers to understand when stricter duties apply, while avoiding retroactive punishment. Finally, by calibrating liability to capacity to influence, the scheme is proportionate to responsibility: those who benefit from heightened authority and attention bear correspondingly greater obligations, whereas ordinary citizens retain the broadest protection for their speech.

Some practical obstacles must be acknowledged. The proposal developed here is primarily normative: it outlines what, in my view, would be justified and how regulation could ideally be organized. Nevertheless, I recognize that certain practical difficulties could hinder its implementation. The criminalization of disinformation would face many of the same challenges as other crimes committed via the internet, such as jurisdictional issues arising from border-dependent criminal law and the general difficulty of enforcement in virtual spaces (see, e.g., Zając Reference Zając2019). All these problems would apply here as well. However, limiting the scope of criminalization to individuals with epistemic power would partly mitigate these obstacles. Those who possess epistemic power are often publicly known, operate under their real names, and frequently monetize their online presence. Compared to anonymous actors or state-sponsored trolls, such individuals would be far easier to identify and hold accountable.