Search the archive.

It took a marathon, 29-hour session in June 2021 for the full House Judiciary Committee to approve a package of six bills aimed at regulating Big Tech. One would have barred Google, Amazon and Apple from steering or requiring customers to use their products and services rather than those of competitors. Another would have limited how Google and Facebook monetize the data they collect from users to earn revenue from advertisers. And another would have prevented tech giants from buying up nascent competitors, as Facebook did with Instagram and Google with Waze and YouTube. And if those efforts proved insufficient to prevent the tech giants from using their dominance in one market to win dominance in related markets, courts would be authorized to break them up.

Yet while each of those bills were approved by bipartisan majorities that spanned the ideological spectrum from Democrat Pramila Jayapal (Wash.) on the left to Republican Matt Gaetz (Fla.) on the right, Buck and Cicilline were already running into resistance.

Despite exonerating the specific defendants for the specific conduct at issue in the two cases, these new doctrines added both unpredictability and complexity to trademark law. In terms of unpredictability, when a circuit court creates a new defense or exception, an attorney can do little better than guess whether another circuit court will adopt it, and if so, in what form. For example, the Second and Third Circuits have adopted the nominative fair use defense, but both have modified it. While the Third Circuit modified the doctrine only slightly,190 the Second modified it far more substantially.191 The Sixth Circuit has so far rejected nominative fair use entirely.192 In terms of complexity, while complexity in trademark rules can improve our ability to match actual to optimal outcomes, complexity also increases the cost of trademark litigation. These increased costs have led, and will continue to lead, to a two-tiered trademark system: one system for those who can afford it, and a second for those who cannot. For those who can afford it, a complex legal system can produce reasonably sensible, albeit expensive, outcomes. For the far larger number of market participants who can no longer afford trademark litigation, theoretically perfect but complex trademark rules offer only theoretical justice. In reality, for those who cannot afford to marshal them, such rules offer no justice at all. Unable to afford the high costs of trademark litigation, these market participants can neither vindicate their own trademark interests when infringed by others, nor defend themselves against excessive and overbroad assertion of trademark rights by well-funded bullies.193

To study the causal effects of varying trademark protection we exploit the Federal Trademark Dilution Act (FTDA) of 1996, which granted additional legal protection to “famous" trademarks until its key provision was nullified in 2003 by a U.S. Supreme Court decision. We find that the FTDA raised treated firms’ operating return on assets by an average of 1.7 percentage points (pp), equal to 12% of their average pretreatment profits. The FTDA was followed by a sharp increase in trademark lawsuits under the new provision and by reduced entry and turnover in affected goods and service classes, consistent with our hypothesis that the FTDA raised the expected cost of entry into affected product markets. Treatment effects were strongest in firms that faced moderate entry threat ex ante, consistent with strategic entry deterrence as a mechanism.

In surveys conducted by the NSF and Census Bureau, both private and public firms rate trademarks as their most important intellectual property assets, ahead of trade secrets, copyrights, and patents (Jankowski 2012). This paper uses the 1996 Federal Trademark Dilution Act to examine the effects of trademark protection on firms' profits and strategy. We document that treated firms' operating profits increased sharply after the FTDA's passage, and fell back after its key provision was nullified by a Supreme Court decision. The value of added trademark protection was localized in industries that had high advertising expenditures, selling expenditures, and trademark holdings ex ante. More affected product markets had sharply lower entry and exit post-FTDA, consistent with a mechanism whereby the FTDA raised entry costs for potential entrants into affected product markets. Treatment effects were strongest in firms that faced moderate entry threat ex ante, consistent with strategic entry deterrence by the incumbent firms. Guided by theory, we investigate the effects of strengthening trademark protection on firm strategy. Treated firms reduced product quality as measured by increased incidence and value of product recalls, and curtailed innovation by reducing R&D spending, patenting, and new product announcements. While these observations are broadly consistent with a quiet life or rent extraction hypothesis, we also find that treated firms introduced new products into previously unexplored markets by extending their brands into all-new goods and service classes. Taken together, these observations suggest that stronger trademark protection led firms to pursue exploitative rather than exploratory innovation and product market strategies. It is important to acknowledge that our results show the effects of the treatment on the treated, that is, the effects of granting additional antidilution protection to incumbent firms in a strong IP enforcement regime. This episode is useful to understand how strengthening the protection of famous trademarks affected firm and industry dynamics. An open question is whether these effects dominate in the long run, and whether there are benefits to strengthening trademark protection for smaller firms and startups or in environments of weak IP enforcement. The comprehensive USPTO trademarks data can contribute to further research on these topics, as well as new research on product market dynamics and intangible capital. We believe these results are an important first step, by demonstrating that firms' profits and strategy are strongly affected by changes in trademark protection at the margin. Our findings that the FTDA was followed by lower product quality and innovation add to recent studies on the potential downside of intellectual property protection (Boldrin and Levine 2002; Williams 2013; Moser 2013), and suggest that strengthening trademark protection may not deliver what is promised on the label.

Next we investigate the relationship between trademark protection and innovation. A large theoretical literature predicts that the incentive to innovate depends on firms' market power, specifically the difference between expected preinnovation and postinnovation rents (Arrow 1962; Aghion and Howitt 1992; Aghion et al. 2009). Both the quiet life hypothesis (Bertrand and Mullainathan 2003) and the Chamberlin (1933) rent extraction hypothesis suggest that stronger trademark protection leads to more managerial slack or more preinnovation rents, and thus to less innovation. Alternatively, the higher profits we document might have relaxed financial constraints and been at least partly plowed back into R&D. Table 4 investigates several measures of firms' innovative activities. Column 1 shows that treated firms reduced R&D spending over assets by 0.32 pp following the FTDA. This is a significant change relative to their pretreatment mean of 3.4%. Column 2 shows that treated firms took out 9% fewer patents post-FTDA. When we weight patents by the adjusted citations they subsequently received, the reduction is greater at —14% (Column 3). However, the estimated dollar value of granted patents for treated firms (from Kogan et al. 2017) actually increased post-treatment by 13% (Column 4). The asymmetry of the citation- and dollar-weighted estimates is striking, and consistent with the FTDA leading treated firms to pursue a less exploratory and more exploitative innovation strategy. Apart from R&D and patents, a common measure of product-level innovation is announcements of major new product lines gathered from newspapers or trade magazines (Chaney, Devinney, and Winer 1991; Mishra and Bhabra 2001). Table 4 Column 5 shows that treated firms were 13% less likely to announce a major new product in the post-treatment period, with the number of new products announced showing a 24% decrease (Column 6). Taken together, these results appear consistent with both the quiet life and trademarks-foster- monopoly hypotheses, or a model in which the FTDA discouraged innovation by raising preinnovation rents. Figure 4 shows graphs of the year by year coefficients for the main innovation outcomes, R&D spending and patents. For brevity the graphs span 1982-2005 to include the TLRA, FTDA, and the Moseley decision. Panel A shows that treated firms' R&D spending was unchanged around the TLRA placebo event, fell after the passage of the FTDA, and reversed to pretreatment levels following the Moseley decision. Panel B shows that new patent grants were unchanged around the TLRA placebo event, fell after the passage of the FTDA, and leveled off following the Moseley decision. These year-by-year coefficient estimates are consistent with the estimates in Table 4 and with our hypotheses that the following the FTDA, treated firms moved from an innovative toward an exploitative product market strategy.

Fourth, trademark law may benefit from the applicability of law developed to address patent trolls.13 For example, the U.S. Supreme Court’s eBay v. MercExchange decision concerning patent injunctions has been applied to trademark injunctions, thus removing a presumption of irreparable harm where a likelihood of success on the merits is demonstrated.14 The U.S. Supreme Court’s Octane Fitness v. Icon Health & Fitness decision concerning attorney fees in patent cases has also been applied in trademark cases.15 Fifth, the litigation and other strategic advantages present in so-called patent trolling cases may not be available in trademark cases.16 Sixth, state attorney general enforcement of state consumer protection laws against patent trolls may deter a trademark troll problem from developing.17 Notably, at least one state has already considered legislation specifically targeted at trademark trolls.18 Moreover, the U.S. Federal Trade Commission’s (FTC) enforcement of federal consumer protection laws against patent trolls may similarly deter trademark trolls.19 Seventh, one commentator has argued that the patent troll problem developed, in part, because of the U.S. Court of Appeals for the Federal Circuit’s (Federal Circuit) nationwide exclusive appellate jurisdiction over patent cases.20 Trademark law is less susceptible to a troll problem developing because the various federal circuit courts of appeals, as well as state courts, have jurisdiction to hear trademark appeals.21 Finally, an early U.S. Court of Appeals for the Seventh Circuit decision addressing trademark trolling upheld an award of attorney fees against a troll and canceled the troll’s mark for lack of use.22 This case has perhaps provided a disincentive to trademark trolling.23

Another event that could stifle the development of trademark trolls is the possibility that state attorneys general will also sue trademark trolls for sending abusive cease-and-desist letters, as has happened with patent trolls.260 Moreover, the FTC recently entered a settlement agreement with a patent troll. The possibility that the FTC may bring an action against trademark trolls may lessen the likelihood that a problem with trademark trolls will develop. The Vermont, Nebraska, and Minnesota Attorneys General have brought suit against patent trolls for the abusive efforts of patent trolls.261 Moreover, the Vermont Legislature passed a law that is specifically targeted at patent trolls.262 The law prohibits “bad faith assertions” of patents.263 The determination of what is bad faith depends on an analysis of various factors, including a failure to include specific information such as specific factual allegations concerning infringement, a failure “to conduct an analysis comparing claims,” whether the claim is meritless or deceptive, and whether the entity asserting the patent has been successful in prior litigation.264 Section 4199 of the legislation provides for attorney fees, costs, and “exemplary damages in an amount equal to $50,000.00 or three times the total of damages, costs, and fees, whichever is greater.”265 Vermont has brought at least one suit against a well-known patent troll under its general consumer protection laws.266 Recently, the FTC reached a settlement agreement with a well-known patent troll.267 Notably, future deceptive conduct will result in relatively substantial fines.268 The use of consumer protection laws, as well as tailored laws, may deter trademark trolls and will certainly influence the behavior of potential trolls.269 7. The Federal Circuit and Uniformity as the Root of Trolls At least one commentator has argued that the Federal Circuit and its case law are essentially responsible for the patent troll problem.270 Other commentators have criticized the Federal Circuit for its supposedly pro-patent approach and argued that the Federal Circuit should share jurisdiction over patent appeals with the other regional circuits.271 Thus, the Federal Circuit’s influence over patent law would be reduced, and the Supreme Court would benefit from having other viewpoints concerning patent law from the circuit courts of appeal.272 Notably, the Supreme Court has heard a number of appeals from the Federal Circuit concerning patent law and has overturned the Federal Circuit numerous times.273 Some of the potential downsides to removing the Federal Circuit’s exclusive jurisdiction over patent appeals would include having a less uniform nationwide patent law274 and the existence of extensive forum shopping. To the extent that the Federal Circuit is responsible for the patent troll phenomena, trademark law should be less susceptible to the development of the same problem because the Federal Circuit shares jurisdiction over trademark appeals275 with the other federal circuit courts of appeal.276 Thus, the problems associated with capture by an interested group277—such as trademark lawyers—are less likely and the Supreme Court is able to get various views from the other circuit courts of appeal concerning trademark issues. Moreover, state courts hear trademark cases as well. Notably, many argue that trademark law has continued to expand; however, some circuit courts of appeal have applied doctrines such as nominative fair use and other doctrines to protect competition and First Amendment values.278 Thus, the nature and structure of the appellate system and development of trademark law could make it less likely that a trademark troll problem could develop similar to the patent troll situation.

The TTAB has jurisdiction over four types of inter partes proceedings: oppositions, cancellations, interferences, and concurrent use proceedings.76 Cancellation proceedings may be initiated by any person who believes that they are or will be damaged by a registered mark.77 An inter partes proceedings before the TTAB is similar to district court proceedings in that there are pleadings, motions, discovery and brief filing.78 They differ, however, because the TTAB is an administrative tribunal and thus, proceedings are conducted solely in writing.79 A party to a proceedings may never even come before the board, by way of giving witness testimony or otherwise, unless a party requests oral hearing on the matter.80

The failure to use a mark in commerce may arise in two contexts: in inter-partes proceedings before the United State Patent and Trademark Office (“USPTO”) and when raised as a collateral attack in litigation. The Lanham Act gives concurrent power to order the cancellation of a mark to the Trademark Trial and Appeal Board (“TTAB”) and the courts.57 In either context, the non-use of a mark may be challenged on grounds of abandonment or a user may allege use of the mark prior to the date of registration thereby creating a priority contest.58 A party may file a petition to cancel a trademark directly with the USPTO subject to certain restrictions as explained supra in Section B.59 One can also challenge the validity of a mark and request a court to order its cancellation during trademark litigation.60

Trademark trolling occurs when an individual or entity registers a trademark in bad faith with no lawful intention of using the said mark. The trolls purposely try to register such marks which have garnered immense reputation and goodwill, in other jurisdictions or classes, which the owner of the mark would have likely want to register in future. Once these marks are registered, they are used as a means for extorting monetary benefits from real mark owners by ways of demanding licensing fees or initiating frivolous infringement suits. In some instances, the real mark owners have to buy their own brand rights from a troll in another country at an exorbitant fee because the troll had first registered the mark in his country. A classic example of trademark trolling is the well-publicized case dating from 2012 when US tech giant Apple Inc. had paid $60 million to Chinese firm Proview to settle a dispute involving rights to the "iPad" name. In 2009, Apple had paid a Proview Technology affiliate in Taiwan for its iPad trademarks in various countries for £35,000, assuming it covered China. However, when the iPad was introduced, Proview’s Shenzhen operation said that the rights were never transferred in China and proceeded to file a trademark infringement case against Apple and prompting authorities to start seizing iPads in February 20121. In another instance of trademark trolling, Tesla Motors Inc., the US electric car maker was forced to settle with Chinese businessman Zhan Baosheng, who had registered the “Tesla” trademark in 2006 much ahead of the automaker’s entry into the Chinese market2. he above instances demonstrate how even large multinational companies have fallen prey to the trademark trolls. These companies had no option but to pay up, as they could not afford to lose access to China’s massive market, which would have severely interrupted their operations and resulted in huge losses. Even celebrities are not immune from the perils of trademark trolling. After almost a decade, U.S. basketball legend Michael Jordan finally secured a major victory in his protracted legal battle against a Chinese sportswear company over the use of his name . Jordan had filed a lawsuit against a sportswear company called Qiaodan Sports (a transliteration of Jordan’s name). The said company had even registered trademarks for the names of his two sons, Jeffrey and Markus (Jiefuli Qiaodan and Makusi Qiaodan).

There are enforcement issues with trademark law and practice. However, trademark law and practice do not suffer from a troll problem for the reasons explained. Those reasons include: the trademark use requirement, Congress’ and ICANN’s response and careful planning to address issues concerning domain names, the availability and widespread use of opposition and cancellation proceedings at the USPTO, Supreme Court precedent addressed to patent trolls applied in trademark cases, litigation and strategic advantages available to patent trolls not present in trademark cases, enforcement of consumer protection laws against patent trolls, shared appellate jurisdiction over trademark cases, and an early appellate case dealing harshly with a trademark troll. If the use requirement and the prohibitions on naked licensing and assignments in gross are discarded or weakened, a trademark troll problem could develop in the United States similar to that in other countries. The problem may not be as severe as the patent troll issue, likely because of the other reasons directed at why a trademark troll problem may not exist.

More than 600 applications have been filed with the U.S. Patent and Trademark Office seeking to register COVID-related marks. And the number of applications continues to grow, with many appearing to be “opportunistic” trademark applications. One, for example, seeks to register “COVID-19 VIRUS FREE GUARANTEED” for apparel and other consumer goods. Another seeks to register “COVIDFREE” for antibacterial goods. Most of these opportunistic applications won’t be approved because, unlike cybersquatting, registration requires actual use or a bona fide intent to use the mark. Most applicants will be unable to establish the requisite intent or simply will never get around to making actual use in commerce. However, for those opportunistic registrants, or potential “trademark trolls,” who have no intent to lawfully use the trademark, the current pandemic presents an occasion to financially exploit these marks. By seeking federal rights ahead of others, the registrant can stand in the way of legitimate businesses making use of these marks. This situation presents an unavoidable business hazard to entities using COVID-related phrases or expressions. With some advance precautions and a nimble response, exposure can often be minimized. After all, the threat posed by most of these trademark trolls is minimal.

The Reasons Why There Is Not a Trademark Troll Problem in the United States This section discusses eight reasons why there is not a trademark troll problem in the United States. The reasons include: the trademark use requirement, Congress’ and ICANN’s response to issues concerning domain names, the availability and widespread use of inter partes administrative proceedings at the USPTO, Supreme Court precedent addressed to patent trolls applied in trademark cases, litigation and strategic advantages available to patent trolls not present in trademark cases, enforcement of consumer protection laws against patent trolls, shared appellate jurisdiction over trademark cases, and an early appellate case dealing harshly with a trademark troll. 1. The Trademark Use Requirement This section discusses the trademark use requirement and how it underlies other doctrines in trademark law which prevent a trademark troll problem from developing in the United States. Those additional doctrines include abandonment, the prohibition against assignments in gross, and the requirement that licensing not be naked.

The assumptions of long-term savings and deterrence as the result of cancellation is flawed for one important reason: because registration merely confers a presumption of the right to use, ownership, and validity, the loss of the registration results in only the loss of a registered right. Because trademark rights arise out of use and not registration, a party may still claim ownership of a valid mark at common law.186 Thus, a party like Stoller, who despite never winning one case and being forced to pay for all the costs of litigation in some, are keen on still filing suit may simply continue to do so. Although the presumptions of validity are not present, one can still claim rights to a mark based on mere use pursuant to common law and state claims of unfair competition. The Lanham Act still provides protection for unregistered marks. As such, the cancellation of a mark does not bar an ambitious trademark troll from bringing suit or threatening to file suit if a licensing agreement is not reached between the parties. Thus, deterrence and resulting long-term savings may be positive thinking rather than a realistic result of proactive cancellation by the courts. However, as discussed below, the threat of not only paying one’s own costs of litigation, but also the other party’s, may provide an even greater push to stop trademark trolls from filing suit for infringement of a mark they do not have rights in.

In “Subjective and Objective Measures of Democratic Backsliding,” the political scientists Andrew Little of the University of California at Berkeley and Anne Meng of the University of Virginia find that the post-Cold War expansion of democracy in nations around the world has, on average, held. “Recent studies that find evidence of global backsliding,” they observe, “rely heavily if not entirely on subjective indicators” — such as asking experts to rate, based on their own judgment, the extent to which an election was “free and fair.”

Little and Meng instead measure democracy based on quantifiable factors less susceptible to individual bias. They find, for example, no overall decline in electoral competition. Across the world, “The rate of leader and ruling party turnover has remained fairly constant since the late 1990s,” the paper says. “If anything, the vote shares of incumbent leaders in executive elections and incumbent parties in legislatures have decreased in recent years.”

Resolution of this controversy is fundamental to the study and practice of international relations. The observation of democratic peace is ‘‘the closest thing we have to an empirical law’’ in the study of global politics (Levy, 1988: 662), and carries the profound implication that the spread of democracy will end war. New economic norms theory, on the other hand, yields the contrary implication that universal democracy will not end war. Instead, it is market-oriented development that creates a culture of contracting, and this culture legitimates democracy within nations and causes peace among them. The policy implications could hardly be more divergent: to end war (and support democracy), the contractualist democracies should promote the economies of nations at risk (Krieger and Meierrieks, 2015; Meierrieks, 2012; Mousseau, 2000, 2009, 2012a, 2013; Nieman, 2015).

In the literature are five factual claims for why we should continue to believe that democracy causes peace: (1) an assertion that in three of the five studies that overturned the democratic peace (Mousseau, 2013; Mousseau et al., 2013a, b), the insignificance of democracy controlling for contractualist economy is due to the treatment of missing data for contractualist economy (Dafoe et al., 2013, henceforth DOR); (2) a claim of error in the measure for conflict (DOR) that appears in one of the five studies that overturned the democratic peace (Mousseau, 2013); (3) an alleged misinterpretation of an interaction term that appears in one of the five studies (Mousseau, 2009) that overturned the democratic peace, along with in inference of democratic causality from an interaction of democracy with contractualist economy (Dafoe and Russett, 2013; DOR); (4) a claim of reverse causality, of democracy causing contractualist economy (Ray, 2013); and (5) a report of multiple regressions with most said to show democratic significance after controlling for contractualist economy (DOR).

We have argued that concerns over the effects of generative AI on the information landscape—and, in particular, the spread of misinformation—are overblown. These concerns are part of an old and broad family of moral panics surrounding new technologies (Jungherr & Schroeder, 2021b; Orben, 2020; Simon & Camargo, 2021). When it comes to new information technologies, such panics might be based on the mistaken assumption that people are gullible, driven in part by the third-person effect (Altay & Acerbi, 2023; Mercier, 2020).

These concerns also tend to overlook the fact that we already owe our current information environment to a complex web of institutions that has allowed the media to provide broadly accurate information, and for the public, in turn, to trust much of the information communicated by the media. These institutions have already evolved to accommodate new media, such as film and photography (Habgood-Cote, 2023; Jurgenson, 2019), even though it has always been possible to manipulate these media. Moreover, it’s far from clear that the more technologically complex forms of manipulation are the most efficient: nowadays, journalists and fact checkers struggle not so much with deepfakes but with visuals taken out of context or with crude manipulations, such as cropping of images or so-called “cheapfakes” (Brennen et al., 2020; Kapoor & Narayanan, 2023; Paris & Donovan, 2019; Weikmann & Lecheler, 2023).

In absolute terms, misinformation already abounds online and, unlike high-quality news or scientific articles, it is rarely behind paywalls. Yet, despite the quantity and accessibility of misinformation, the average internet user consumes very little of it (e.g., Allen et al., 2020; for review, see Acerbi et al., 2022). Instead, misinformation consumption is heavily concentrated in a small portion of very active and vocal users (Grinberg et al. 2019). What makes misinformation consumers special is not that they have privileged access to misinformation but traits that make them more likely to seek out misinformation (Broniatowski et al., 2023; Motta et al., 2023), such as having low trust in institutions or being strong partisans (Osmundsen et al., 2021). Experts on misinformation view partisanship and identity as key determinants of misinformation belief and sharing, while they believe lack of access to reliable information only plays a negligible role (Altay et al., 2023). The problem is not that people do not have access to high-quality information but instead that they reject high-quality information and favor misinformation. Similarly, conspiracy theories exist everywhere and are easily accessible online across the globe. Yet, despite similarities in supply, demand for conspiracy theories varies across countries, such that in more corrupt countries, conspiracy theories are more popular (Alper, 2023; Cordonier & Cafiero, 2023).

Finally, (mis)information, on its own, has no causal effect on the world. (Mis)information only gains causal powers when humans see it. Yet, the number of things that go viral on the internet and get seen is finite because our attention is finite (Jungherr & Schroeder, 2021a; Taylor, 2014). And since generative AI is unlikely to increase demand for misinformation and will not increase the number of things humans can pay attention to, the increase in misinformation supply will likely have limited influence on the diffusion of misinformation.

Methodologically, it is very difficult to research whether fake news indeed alters voting behaviour (Allcott & Gentzkow, 2017).1 While ‘fake news’ indeed spreads fast around social networks, this does not mean that people actually believe these stories or that they change actual behaviour or political convictions. It is true that disinformation seems to have at least some power to set the agenda of mainstream news (Vargo et al., 2017). But even though confirmation bias and polarization play a role (Del Vicario et al., 2016), echo chambers do not play a decisive role in the outcomes of elections (Bail et al., 2020; Boulianne et al., 2020). Research on the spread and impact of fake news usually assumes that people sharing these stories believe in them and that reading these stories will alter their political behaviour. However, sharing stories is never only about sharing (neutral) information but also involves an affective dimension (Wardle & Derakhshan, 2017, p. 7). Based on previous research, we can detect two ways in which fake news is ultimately not about the (lack of) veracity of the stories but the way they work on an affectual dimension. The first crucial aspect concerns clickbait, the second how sharing stories is an expression of identity. Let me briefly discuss both aspects in turn. As Tommaso Venturini states succinctly ‘junk news is not about algorithmic persuasion’ (Venturini, 2019, p. 123). He, among others, argues that people spread ‘fake news’ items even though they do not believe them. They are shared because people might want to debunk them or make fun of them (Venturini, 2019, p. 126). Since the point is not about the ‘fakeness’ as such, Venturini prefers to speak about ‘junk news’ or ‘viral news’ (Venturini, 2019, p. 126). Fake news items work since they are addictive and grab our attention – even if we do not believe them. They are produced because they hold our attention which can then be translated into monetary revenue. A whole political economy has developed around fake or junk news (Bakir & McStay, 2018). This shows that fake news as a phenomenon is very complex and we need to consider economic aspects that shape how news stories are produced and shared. Indeed, when we look at research on fake news in the field of sociology, we can observe another reason why the phenomenon of fake news is not best explained by only looking at quantitative signifiers. Following up on the role of the affective dimension, Polletta and Callahan (Polletta & Callahan, 2017; see also: Hochschild, 2016b) have proposed that we should understand the sharing of fake news less as an act of rational information sharing than as an act of identity expression. The authors argue that sharing fake news in the US presidential election can be better understood as an expression of a common identity. The question of whether people ‘really’ believed in these stories is then secondary. Disinformation thus becomes something that is not simply about a lack of facts in news but also about people being emotionally attached to certain news and their emotional reactions that help them assess the veracity of news stories (Boler & Davis 2018, pp. 78–79). People identify with certain news and use the sharing of news as an expression of their identity and their belonging to a specific social group (Polletta & Callahan, 2017, p. 400). We should be less concerned with the direct impact of disinformation than with what it reveals as a symptom of broader changes in political communication and media consumption. Junk news aim at affective reactions such as a feeling of partisanship or deep social division (Venturini, 2019; Young, 2021, p. 5). Sharing fake news seems to occur for multiple reasons and the emergence of ‘information disorder’ or ‘junk news’ is a symptom of changes in the media landscape and a distinct political economy based on clickbait. As a consequence, more recent research has investigated the debates that revolve around fake news (Creech, 2020; Farkas & Schou, 2018; Monsees, 2020; Wasserman, 2020). That is why, in this article, I focus on how debates over fake news and disinformation rely on specific arguments and ideas that ultimately help, for example, delegitimise certain opinions or present fake news as a security threat. Conceptually, I read the debates on fake news

in Germany, Czechia and the US through ideas on democracy. In the next section, I discuss democratic theories and their presence in debates on disinformation.

First, it seems that it would already be relatively easy for producers of misinformation to increase the perceived reliability of their content. In the realm of visuals, Photoshop has long afforded bad actors the ability to make an artificially created image look real (Kapoor & Narayanan, 2023). If they opt not to do so, it might be because making a text or image look more reliable or real might conflict with other goals, such as making the same more accessible, appealing, or seemingly authentic. It’s not clear that generative AIs could increase content quality on multiple dimensions, as there might be some unavoidable tradeoffs.

Second, even if generative AI managed to increase the overall appeal of misinformation, most people are simply not exposed, or only very marginally exposed, to misinformation (see above and Acerbi et al., 2022). Instead, most people overwhelmingly consume content from mainstream sources, typically the same handful of popular media outlets (Altay et al., 2022; Guess et al., 2021). As a result, any increase in the quality of misleading content would be largely invisible to most of the public.