The common law method is to try to apply existing precedents to new situations. It is therefore to be expected that lawyers and judges would apply well-established legal principles to the relatively new internet and social media. Although some basic principles apply to these media, as they do to all others, there are ways in which this approach fails because the internet and social media are different in kind from other methods of communication. Scholars and legal experts have given insufficient attention to the areas in which traditional First Amendment principles are inadequate for dealing with the internet and social media.

In this chapter, I begin by describing how the internet and social media are inherently different from other media. I then discuss traditional First Amendment doctrines that should apply to these newer media. With this as context, I then focus on three areas where current doctrine is inadequate for dealing with the internet and social media: These are false speech, how the internet and social media undermine the traditional press, and the issue of foreign influence in elections. My goal is not to offer solutions but to identify these areas where applying existing approaches is likely to fail because of the unique nature of the internet and social media as places for communication.

6.1 How the Internet and Social Media Are Different from Other Forms of Communication

How do the internet and social media differ from other forms of communication? First, the internet has democratized the ability to reach a mass audience. It used to be that to reach a large audience, a person had to be rich enough to own a newspaper or obtain a broadcast license. Now, though, anyone with a smartphone – or even just access to a library with a modem – can reach a huge audience instantaneously. No longer are people dependent on a relatively small number of news sources.

Scarcity of media outlets was a core characteristic of media until now. There are only so many broadcast channels,Footnote ¹ and there are only so many newspapers that the economy will support, with each having only so many physical pages. But the internet and social media provide the opportunity for infinite speech. Scarcity is no longer an issue.

Further, the internet and social media empower individuals who would otherwise be voiceless to express their opinions and provide them a place to do so.Footnote ² The internet provides a platform for those who cannot physically gather and organize.Footnote ³ Previously silenced or marginalized individuals now have an accessible method for disseminating their political ideas.Footnote ⁴

But these characteristics of the internet and social media also mean that false information can be quickly spread by an almost infinite number of sources. Information that is true but confidential can be quickly disseminated.Footnote ⁵ There is even a name for such action: “doxing,” which refers to publishing private information about a person on the internet, often with the malicious intent to harm the individual.Footnote ⁶ The internet and social media can be used to harass.Footnote ⁷ A study by the Pew Research Center “found [that] 40 percent of adult Internet users have experienced harassment online, with young women enduring particularly severe forms of it.”Footnote ⁸

Traditionally, newspapers have acted as an important filter by excluding false and other types of harmful information. They have an editorial process and, if not journalistic ethics, at least a fear of liability. Certainly, at times, the filters have failed and newspapers have done harm. But at least there are filters, while none exist for the internet and social media. Anyone can post anything for the world to see.

Second, the internet has dramatically increased the dissemination and permanence of information, or to phrase this differently, it has enormously increased the ability to access information. Lawyers and law students can do all of their research online, reading statutes, cases, and treatises – previously, this would have required a trip to the law library. We can visit the great museums of the world online anywhere, anytime. We have access to virtually unlimited information from nearly infinite sources.

But there is a downside to this easy access and permanence. Take defamation as an example. Imagine, before the internet, that a local newspaper published false information about a person that harmed their reputation. The falsity would be known by readers of the paper and could be circulated by other media and by word of mouth. Eventually, however, the impact of the defamatory article would fade, and the false information would live on only in remote places like a library’s microfiche archives. Now, though, the defamatory story can be quickly spread across the internet, accessed from anywhere, and is likely to remain there forever. It is enormously difficult, if not impossible, to erase something from the internet.

Finally, the internet does not respect national boundaries. Again, there are great benefits to this – for instance, totalitarian governments cannot easily cut off information to their citizens. When the revolution began in Egypt, the government tried to halt internet access, but people with satellite phones maintained access and disseminated what they learned.Footnote ⁹ Yet, there are also drawbacks. The Supreme Court has estimated that 40 percent of pornography on the internet comes from foreign countries, making any attempt to control it within a country impossible.Footnote ¹⁰ As occurred in the 2016 presidential election and evidenced by Special Counsel Robert Mueller’s report, the lack of borders on the internet also gives foreign countries and foreign actors a vehicle for trying to influence the outcome of United States elections.Footnote ¹¹

6.2 Basic First Amendment Principles Should Apply to These Media

I do not want to overstate the First Amendment problems in dealing with newer media. The same basic First Amendment principles should apply to these media as to all other aspects of the press. Central to free speech analysis is whether a government regulation is content-based or content-neutral, with content-based regulations having to meet strict scrutiny and content-neutral laws needing to satisfy intermediate scrutiny. As the Court observed in Police Department of Chicago v. Mosley, “[A]bove all else, the First Amendment means that government has no power to restrict expression because of its message, its ideas, its subject matter or its content.”Footnote ¹² Hence, the Court endorsed a two-tier system of review. The Court applies “the most exacting scrutiny to regulations that suppress, disadvantage, or impose differential burdens upon speech because of its content.”Footnote ¹³ But, “[i]n contrast, regulations that are unrelated to the content of speech are subject to an intermediate level of scrutiny.”Footnote ¹⁴

This should apply to government regulation of the internet and social media as it does to all other government regulation of speech. Likewise, the same categories of unprotected speech – incitement, true threats, defamation, false and deceptive advertisements – are applicable to expression made over the internet and social media.

A core aspect of freedom of the press is that the government cannot control the content of what is published. This should apply with equal force to the internet and social media. In Miami Herald v. Tornillo, the Court unanimously declared a right-of-reply law unconstitutional as applied to newspapers.Footnote ¹⁵ A Florida law required that a newspaper print a reply from any candidate for office whose character or official record had been attacked in its pages. The newspaper had to print the reply free of charge and in as conspicuous a place as the initial story was presented. The Court, in an opinion by Chief Justice Burger, stated the following:

The Court stressed that forcing newspapers to publish a reply intrudes on editorial discretion that is protected by the First Amendment. Likewise, regulating what social media companies do or don’t “publish” should be deemed to violate the First Amendment.Footnote ¹⁷

6.3 The Internet and Social Media Raise Unique Issues

Although the fundamental principles of the First Amendment apply to the internet and social media, it cannot be ignored that these newer media pose unique issues for which existing doctrine does not provide an adequate answer. First, there is the problem of false speech. False information can cause great harm, and social media facilitates the rapid dissemination of such information.Footnote ¹⁸ The unprecedented speed and scale of digital communication have simplified the dissemination of false information like never before, frequently outpacing efforts to debunk it.Footnote ¹⁹ Social media algorithms can prioritize content more likely to generate user engagement regardless of its veracity.Footnote ²⁰ Facebook and other social media are sending their users specific content based on the platform’s assessment of their interests but with little regard to whether the information is true. In the political realm, this can change the outcome of elections, especially if the falsehoods are spread soon before election day when there is little time to refute them. The spread of disinformation and misinformation can prevent voters from accessing essential information required to make well-informed decisions regarding candidates for office and laws.Footnote ²¹ False accusations can be quickly and widely circulated, making all the difference between who wins and who loses. False speech has enormous implications for politics, public health, elections, and other domains where the availability of accurate information is vital. Additionally, and quite importantly, false information can undermine trust in reliable sources and credible news.Footnote ²²

Researchers at the Massachusetts Institute of Technology “found that false news spreads faster and to more people than true stories, reaching more people than any other type of information.”Footnote ²³ The First Amendment is predicated on the romantic notion – articulated by John Stuart Mill and Oliver Wendell Holmes – that there is a marketplace of ideas and that allowing all speech will lead to truth triumphing over falsehoods. But many studies have demonstrated that people are unlikely to change false beliefs when confronted with true information.Footnote ²⁴ False information has a permanent effect. When 70 percent of Republican voters believe that Trump was cheated out of the presidency in 2020, how can there be faith in the notion that truth will triumph over falsity in the world of the internet and social media?

This problem will grow as technology becomes more sophisticated. It is all too easy to imagine how, in the days before an election, unscrupulous actors could circulate a deepfake of an opposing candidate saying something repulsive. What we see and hear for ourselves seems true, even if it is created by bots or artificial intelligence. A deepfake depicting a politician taking a bribe would create lingering doubts for many, and be believed by some, even after it is revealed to be fictitious.

In fact, one place where the law has changed is in the area of injunctions in defamation cases. Traditionally, the only remedy in defamation cases was money damages, in accord with the maxim that equity would not enjoin defamation.Footnote ²⁵ But because of the internet and social media, the law in this area has changed, and injunctions in defamation cases have now become commonplace.Footnote ²⁶

A solution to false speech over the internet and social media is likely elusive under existing First Amendment doctrine. The Supreme Court has rightly and emphatically declared the importance of protecting false speech. The most important case in this regard – and one of the most important free speech decisions of all time – is New York Times Co. v. Sullivan.Footnote ²⁷ The Court reversed a defamation judgment against the newspaper and stated that the fact that some of the statements were false was not sufficient to deny the speech First Amendment protection.Footnote ²⁸ The Court explained that a false “statement is inevitable in free debate and [it] must be protected if the freedoms of expression are to have the ‘breathing space’ that they ‘need … to survive.’”Footnote ²⁹ This surely is correct. If any false statement about a government official, no matter how minor, would be a basis for liability, speech about our government would be chilled and lost.

Subsequently, in a very different context, the Court again recognized the importance of judicial protection of false speech in United States v. Alvarez.Footnote ³⁰ A federal law made it a crime for a person to falsely claim to have received military honors or decorations.Footnote ³¹ Writing for a plurality of the Court, Justice Kennedy expressly rejected the government’s argument that false speech is inherently outside the scope of the First Amendment.Footnote ³² He declared that there is not “any general exception to the First Amendment for false statements” and stressed that “some false statements are inevitable if there is to be an open and vigorous expression of views in public and private conversation, expression the First Amendment seeks to guarantee.”Footnote ³³ Justice Kennedy further explained that “the Court has been careful to instruct that falsity alone may not suffice to bring … speech outside the First Amendment.”Footnote ³⁴

Given its harms, why is there First Amendment protection for false speech? In large part, it is because of the danger of giving anyone, including (or maybe especially) the government, the power to decide what is true and what is false. Allowing the government to prohibit false speech grants it the role of arbiter of truth. As Justice Kennedy explained: “Our constitutional tradition stands against the idea that we need Oceania’s Ministry of Truth.”Footnote ³⁵ Would Democrats want the Trump administration, or would Republicans want the Biden administration, to have the power to create an agency that could scour the internet and social media and remove what it deems false? It is so much easier to put faith in the marketplace of ideas to refute falsehoods, even when we know that it is likely to fail in many instances.

But it is wrong to think that the First Amendment never allows restrictions of false speech. There are contexts in which the Supreme Court has refused to protect false speech and has allowed it to be prohibited and punished. For example, it is clearly established that the First Amendment does not protect false and deceptive advertisements.Footnote ³⁶ The government, of course, can constitutionally prohibit making false statements under oath (perjury) or to law enforcement officials. But these are limited areas where false speech can be punished. The Court generally has treated commercial speech as being of lower value than political speech, which may make it easier to say that false advertising is entitled to no constitutional protection. As for perjury, there must be proof of a knowing and intentional falsehood.

The Court’s inconsistent statements about false speech can be understood as reflecting the competing interests inherent in First Amendment analysis. On the one hand, false speech can create harm, great harm even. Speech is protected particularly because of its importance for the democratic process, but false speech can distort that process. Speech is safeguarded, too, because of the belief that the marketplace of ideas is the best way for truth to emerge. But false speech can infect that marketplace, and there is no reason to believe that truth will triumph. The Supreme Court has recognized this and declared that “[f]alse statements of fact are particularly valueless [because] they interfere with the truth-seeking function of the marketplace of ideas.”Footnote ³⁷

A great deal of the problem is simply the amount of speech transmitted over social media and how difficult that makes content moderation. The quantity of information posted daily is staggering, and social media companies already conduct an enormous amount of content moderation. For example, Facebook reported that between October and December 2021, it took action against terrorism content 7.7 million times, bullying and harassment 8.2 million times, and child sexual exploitation material 19.8 million times.Footnote ³⁸ In the last quarter of 2020, Facebook took action on over 1.1 million pieces of content per day.Footnote ³⁹ Designing a system to exclude false speech – even if it could be decided what is false and even if mandating this would not violate the First Amendment – seems an insurmountable task.

But one should have no illusions. There will be elections decided because of false speech circulated over social media, especially in the days before voting. The threat to democracy is real and will grow. We need new First Amendment doctrines to deal with false speech over the internet and social media.

In addition, we must acknowledge how the internet and social media endanger traditional media and thus undermine the crucial newsgathering functions of the press. Much of the country is now facing a crisis of news deserts, which refers to places where local newspapers no longer exist.Footnote ⁴⁰

A recent study by the Medill School of Journalism found that since 2005, the country has lost more than one-fourth of its newspapers and is on track to lose one-third by 2025. Between late 2019 and May 2022, more than 360 newspapers closed. The report concluded:

Most newspapers that have survived have had to cut their staffs. From 2008 to 2017, total newspaper newsroom staffing almost halved, dropping from 71,000 employees to 39,000.Footnote ⁴²

Many studies have shown that when a community does not have a strong print or digital news organization, corruption increases and voter participation decreases. One study of eleven California cities found that when there are fewer reporters covering an area, fewer people run for mayor and fewer people vote.Footnote ⁴³

The internet and social media have contributed to this because they allow people to get their news for free, and thus, they do not need to purchase newspapers. Few newspapers have successfully found a way to monetize their online presence. Google, Facebook, Instagram, and others link to news stories and reap the advertising revenue, with none of the money going to the news organizations that did the actual reporting. The tech platforms are swallowing the profits that otherwise would go to the newspapers that are covering events, conducting investigations, and writing stories.

When news organizations vanish, there is no one to go to the city council meetings or do the investigative reporting to uncover corruption or report on local elections. Google and Facebook and Instagram do not have newsrooms to do this work, and they are not going to create them.

There is a bill pending in the California legislature – the California Journalism Preservation Act – that would help solve this problem. Assembly Bill 886 would require platforms to pay a fee for the news on their platforms, which would allow journalism providers to recoup their fair share of revenue for the news stories they produce. The law would require that these internet and social media platforms pay a “journalism usage fee” to news outlets for news content appearing on their sites. The law would create an arbitration process with newsrooms and establish a fee that the companies would pay to carry news articles.

The experience of other countries suggests that this type of legislation works. In 2021, Australia passed a law requiring social media companies to share a portion of the profits they make from news content with news companies. The law generated almost $150 million in revenue in its first year.Footnote ⁴⁴ Canada adopted a similar law last year.

Yet, it is unclear whether such targeted financial assessments would survive First Amendment scrutiny. In Arkansas Writers’ Project, Inc. v. Ragland, the Court ruled that the government cannot discriminate among types of publications.Footnote ⁴⁵ A state exempted from its sales tax special interest publications, such as religious, professional, trade, and sports journals, but did not exempt general interest magazines. The Court emphasized that any differential taxation of the press – either of the press as opposed to others in society or of particular parts of the press – risked chilling reporting. Yet there is a key distinction: The current proposal would not impose a tax on social media but would rather require that they pay for what they use.

The point is that the internet and social media endanger the press, and protecting freedom of the press requires deliberate action. First Amendment doctrines will need to be adapted to accomplish this.

Finally, another way the internet and social media threaten democracy is when foreign nations and foreign entities use them to influence the outcome of elections. As mentioned earlier, there is now incontrovertible evidence that Russia engaged in a concerted effort to use speech, including false speech, to influence the outcome of the 2016 presidential election.Footnote ⁴⁶ This was suspected during the campaign, and American intelligence agencies confirmed it soon after the election.Footnote ⁴⁷ In February 2018, Special Counsel Robert Mueller issued a 37-page indictment charging thirteen Russians and three companies with executing a scheme to subvert the 2016 election and help to elect Donald Trump as president.Footnote ⁴⁸ Mueller’s indictment details “how the Russians repeatedly turned to Facebook and Instagram, often using stolen identities to pose as Americans, to sow discord among the electorate by creating Facebook groups, distributing divisive ads and posting inflammatory images.”Footnote ⁴⁹

Although condemning Russia’s meddling in the American election is easy, solving the underlying First Amendment issue is difficult. Certainly, illegal conduct, such as hacking into the Democratic National Committee headquarters and subsequently disseminating unlawfully gained information,Footnote ⁵⁰ is not constitutionally protected. But what about foreign speech that is legal and that expresses an opinion – or even false speech?

The Supreme Court has repeatedly stated that the source of information does not matter for First Amendment purposes. In First National Bank of Boston v. Bellotti, in 1978, the Supreme Court declared unconstitutional a Massachusetts law that prohibited banks or businesses from making contributions or expenditures in connection with ballot initiatives and referenda.Footnote ⁵¹ Justice Powell, writing for the Court, concluded that the value of speech is in informing the audience. Any restriction on speech, regardless of its source, therefore undermines the First Amendment. Justice Powell explained: “The inherent worth of the speech in terms of its capacity for informing the public does not depend upon the identity of its source, whether corporation, association, union, or individual.”Footnote ⁵²

The Court relied heavily on this in Citizens United v. Federal Election Commission to hold that corporations have the constitutional right to spend unlimited amounts of money directly from their treasuries to elect or defeat candidates for political office.Footnote ⁵³ The Court stressed that the value of the speech does not depend on the speaker’s identity and held that corporate speech is protected not because of the inherent rights of corporations but because all expression contributes to the marketplace of ideas. The Court wrote that “the First Amendment bars regulatory distinctions based on a speaker’s identity.”Footnote ⁵⁴ On other occasions, too, the Court has declared that “[t]he identity of the speaker is not decisive in determining whether speech is protected.”Footnote ⁵⁵

But if this is so, why should it matter that a speaker is a foreign government or a foreign individual? Federal law prohibits foreign governments, individuals, and corporations from contributing money to candidates for federal office.Footnote ⁵⁶ A federal court upheld this restriction on foreign speech, declaring, “It is fundamental to the definition of our national political community that foreign citizens do not have a constitutional right to participate in, and thus may be excluded from, activities of democratic self-government.” As a result, it asserted that “the United States has a compelling interest for purposes of First Amendment analysis in limiting the participation of foreign citizens in activities of American democratic self-government, and in thereby preventing foreign influence over the US political process.”Footnote ⁵⁷ But can this be reconciled with the Supreme Court’s declaration that the speaker’s identity should not matter in First Amendment analysis? Although it is not a comfortable answer, I do not see a way to exclude foreign speakers that would be consistent with the Court’s premise that the speaker’s identity cannot be the basis for regulation. The First Amendment assumes that more speech is better, whether the speaker is foreign or domestic.

At the very least, however, it would be desirable to disclose speakers’ identities so that people can know when speech is coming from a foreign government or other foreign source. But this, too, raises First Amendment issues, as the Supreme Court has held that there is a First Amendment right to speak anonymously. In Mclntyre v. Ohio Elections Commission, the Court declared unconstitutional a law prohibiting the distribution of anonymous campaign literature.Footnote ⁵⁸ Justice Stevens, writing for the Court, stated that “an author’s decision to remain anonymous, like other decisions concerning omissions or additions to the content of a publication, is an aspect of the freedom of speech protected by the First Amendment.”Footnote ⁵⁹ Further, Justice Stevens contended that anonymity also provides a way for a speaker “who may be personally unpopular to ensure that readers will not prejudge her message simply because they do not like its proponent.”Footnote ⁶⁰

All that said, there is a compelling interest in stopping speakers from masking their identity and deceiving voters. It is one thing to speak anonymously but another to falsely present oneself, especially to deceive and manipulate voters. Whether the speaker is domestic or foreign, that should be regarded as a form of fraud that is unprotected by the First Amendment. At the very least, we should implement laws that prohibit this, along with stronger disclosure laws, especially for foreign speakers. The Supreme Court has consistently upheld the constitutionality of laws that require disclosure of the identity of those spending money in election campaigns. These requirements must be significantly strengthened.

The transnational nature of the internet makes exercising control over speech elusive, even if it were to be constitutional. As the 2016 presidential election demonstrated, foreign governments can use the internet and social media to influence elections without their officials and agents ever physically entering the United States. It is unclear how American law could be successfully applied to them. The internet thus enables them to engage in false speech (and all other kinds of expression) with relatively little fear of legal sanctions, and whether there could be meaningful international sanctions is uncertain at best.

6.4 Conclusion

It is crucial to recognize that the internet and social media are different from the modes of communication that preceded them and that they will require new doctrines for First Amendment analysis because of the inadequacy of the doctrines applied to traditional media. The internet and social media provide great benefits in terms of speech, but they also pose unprecedented harms as well.

The idealized press, which generated effusive praise from the Supreme Court during the golden age of media law in the late twentieth century, was a mature, economically robust, and politically powerful industry. As an industry, the press in that era fulfilled a number of key functions, such as setting the agenda of public issues, swaying the conduct of political campaigns and elections, informing the public of current events, and curating a common understanding of newsworthy developments. The news media had a comfortable relationship of give-and-take with the government but also acted as the Fourth Estate by uncovering, publishing, and amplifying information about government malfeasanceFootnote ¹ and political disinformation.Footnote ² Against this backdrop, courts praised newspapers for their multiple functions as public proxies, audience educators, and watchdogs tasked with keeping government abuse in check. In turn, surveys indicated that the public had faith in the press.Footnote ³

But over the last several decades, a confluence of economic, technological, and sociopolitical change has upended both the news industry and the public’s view of it. While it is unrealistic to hope for a return to the old days of the press, the functions the press has historically played remain essential to democracy. Indeed, one can imagine the press playing an even more robust role today in combating disinformation disseminated by a variety of sources. Yet the news industry today is no longer economically powerful: Newspapers are in peril, television and cable viewership are in decline, news deserts dot the landscape, and jaw-dropping numbers of journalistic staff cuts have drained expertise from newsrooms.Footnote ⁴ Social media have cannibalized content and replaced a scarcity of frequencies with a scarcity of attention, the news industry has far less political power, cell phones have allowed everyone to claim citizen journalist status, both the right and the left have complaints about the asserted bias of press organs, and a common understanding of public issues is a pipe dream in a highly polarized political landscape.

Under these circumstances, the fate of the press’s functions is an existential question both for the news media as we know it and for contemporary American democracy. Managing the complexity of this kind of information environmentFootnote ⁵ calls for independent and principled engagement with issues of public concern by those who hew to journalistic values, such as truth, verification, completeness, investigation, and context. It also calls for appropriate legal protections.

In this chapter, I seek to explore the evolving mosaic of threats facing the American press and consider what, if any, legal rights the wielders of the press function need in response. I begin by identifying a set of key threats facing the press – economic, legal, technological, and audience-based developments. I then propose some initial responses to these threats along five dimensions – funding conditions, a mixed legal strategy, artificial intelligence (AI) policy, industry restructuring, and trust enhancement. I also call for a commitment to press self-examination from the vantage point of fundamental journalistic values in a democracy.

7.1 Evolving Threats to Press Functions

Challenges to press functions today can be grouped into four categories – institutional/journalistic, legal, technological, and audience-based. In combination, they undermine journalistic norms and routines, erode the perceived legitimacy of the press, and hamper its democracy-promoting work.

7.2 The Multiplicity of Responses Needed

These various challenges to the functions of the press cannot be solved by interventions in any one area alone. The particular sorts of threats identified should influence the particular responses, and this should be an iterative and evolutionary process.

7.3 Conclusion

An effective press sector consisting of a variety of actors committed to journalistic values can support democracy. It can know, inform, and educate its many publics. It can discover what the powerful do not want discovered, act as a watchdog, dispel disinformation, and help sustain local and underrepresented communities. Shoring up these functions for the press is a worthwhile enterprise in the public interest. But today’s economic, technological, legal, professional, and consumer contexts are a far cry from those of the press’s golden age. Still, the press and its allies should continue to proselytize the value of journalism’s ideal functions – while acknowledging the instances in which the modern press fails to meet the mark. They should pressure-test reform proposals for viability; develop responsible technology policies; revisit antiquated functions and industry expectations; and use the lessons of interdisciplinary research to explore effective ways to counter public distrust. They should also call for normalizing self-examination across the entire sector to enhance compliance with fundamental journalistic norms. Such targeted steps can help rehabilitate journalism’s public role and allow society to reap the benefits of an independent press. We ought not surrender this promise.

Just a half-century ago, “the news media were one of America’s most trusted institutions.”Footnote ¹ Today, however, almost half of Americans say they have little or no trust in newspapers, and more than half say the same about television news.Footnote ² Although some degree of public distrust of the media is healthy, distrust of this magnitude represents a crisis of legitimacy.

The role of the press in safeguarding democracy is enshrined in the U.S. Constitution. The Constitution protects press freedom not as an end in itself and “not as a favor to a particular industry”Footnote ³ but as a means of making our system of government possible. The framers recognized that the republic’s fate depends on citizens having access to reliable and timely information about their world. Trustworthy news enables us to participate in informed public discussion, forge consensus, and hold democratically elected leaders accountable.Footnote ⁴ If we lack access to (or cannot agree upon) the most basic facts about global, national, and local events or the myriad issues that affect our lives, it is impossible to understand and respond to the challenges facing us and our country. Without shared facts about the world, the stability of our communal life is in jeopardy.Footnote ⁵

It may seem quixotic to lament the public’s loss of faith in traditional media actors at a time when we have – on demand – an unprecedented level of access to information.Footnote ⁶ Social media enables more people than ever before to share content and commentary with mass audiences. Yet this very informational abundance – soon to be magnified by generative artificial intelligence – can make it increasingly difficult to know which information to trust. And even though some types of information are superabundant, other types are becoming scarcer. Case in point: Historically, local newsgatherers played a key, democracy-enhancing role by keeping their communities informed about local events and holding local elected officials to account. Now, however, the market for local news has evaporated, turning more and more cities into “news deserts.” Meanwhile, fewer national legacy news providers can afford to invest in the processes and expertise needed to produce high-quality news about our increasingly complex world.Footnote ⁷ As such, the true crisis of press legitimacy is the declining cultural investment in the systematic gathering of high-quality news produced by independent, transparent, and trustworthy sources.

What caused the Fourth Estate to crumble? Although scholars usually point to a handful of cultural and economic factors as undermining news quality and press credibility, various critics now identify a more covert culprit: the U.S. Supreme Court. The Court is partly to blame for the press’s declining credibility, these critics claim, because the Court’s First Amendment decisions hinder the ability of state defamation law to hold the press accountable for defamatory falsehoods. The implication is that the press would regain much of its credibility if the Court were to remove these constitutional barriers – especially the requirement that public officials and public figures demonstrate “actual malice” on the part of the press for a defamation claim to prevail. As this chapter explains, the current landscape of high-profile defamation cases, and the public reaction to them, casts doubt on whether restoring public trust in the press could be so easy.

8.1 The Usual Suspects

Scholars and commentators have identified a number of potential contributors to the deterioration of public trust in the press. For example, many observers suggest that the growing distrust in the news media is simply a facet of the public’s declining trust in all types of public and private institutions.Footnote ⁸ A recent Gallup poll found historically low levels of public trust in the justice system, the medical system, the presidency, public schools, Congress, and many other institutions.Footnote ⁹ However, even in the context of this overall decline of institutional trust, the growing skepticism toward news media is exceptional: Journalists are less trusted than most other professionals, and only 11 percent of those polled expressed a great deal or quite a lot of confidence in television news; newspapers fared slightly better, but only 16 percent of those polled expressed confidence in them.Footnote ¹⁰

A primary contributor to public distrust in traditional news media providers is the collapse of the economic model supporting news production. Over the past quarter of a century, the Internet has gobbled up advertising revenues that were once used to finance news production, while simultaneously subjecting traditional news providers to increased competition from low-quality “cheap news.” The rise of generative artificial intelligence threatens to accelerate these competitive forces,Footnote ¹¹ which fuel public distrust in the press by making fact-based news both less profitable to produce and harder to distinguish from misinformation. These forces also change both the style and substance of news content in ways that can undermine press credibility.

Instead of relying on traditional media to set their news agenda, many Americans now rely on social media connections, algorithms, and influencers to curate their (mostly free) news content.Footnote ¹² Although some of this free content is produced by news entities that once would have reaped advertising revenue from providing it, the vast bulk of that revenue is now rechanneled to online classified advertising sitesFootnote ¹³ or to platforms that offer the kind of micro-targeted advertising precision that even digital news providers simply cannot match.Footnote ¹⁴ This flight of advertising revenue has disrupted various legacy news producers,Footnote ¹⁵ but newspapers – long the bedrock of news production and resource-intensive investigative journalismFootnote ¹⁶ – have experienced the most dramatic impact.Footnote ¹⁷

It is not just advertisers who have fled to the tech platforms – so, too, have audiences.Footnote ¹⁸ Between roughly 2000 and 2022, daily newspapers lost two-thirds of their readers,Footnote ¹⁹ with total revenues falling by over 80 percent.Footnote ²⁰ While The New York Times and The Wall Street Journal have been able to offset lost advertising revenue with subscription dollars,Footnote ²¹ they are the rare exceptions. Between 2004 and 2020, more than 2,100 papers closed,Footnote ²² and some of the survivors dwindled to become what the NewsGuild has labeled “ghost newspapers” – downsized shells of their former selves.Footnote ²³

The declining financial fortunes of newspapers have affected overall news quality across the media landscape. How could they not? The brain drain in journalism over the last two decades has been cataclysmic. Between 2008 and 2020, newspapers cut newsroom employees by 57 percent.Footnote ²⁴ In 2023, newsrooms experienced the “worst job cuts” in years, and 2024 began with reports of storied newspapers – including, for example, the Los Angeles Times and the Washington Post – delivering more bad news about newsroom staffing.Footnote ²⁵ However, it is not only the evisceration of expertise that hurts news quality. It is also that resources have been disproportionately withdrawn from the production of local news and investigative journalism; these are the very kinds of journalism that help citizens participate meaningfully in shaping their communities and holding officials to account. Without journalistic watchdogs, corruption flourishes and communities suffer.

In the meantime, the need to compete for attention in a crowded marketplace has shaped news content in ways that magnify public distrust. To survive, legacy media have been forced to compete to break news faster than their online competitors, to hijack readers’ attention with sensational content, and, in many cases, to cultivate niche audiences by appealing to partisan loyalties.Footnote ²⁶ This competition forces the legacy press to try to give audiences what they want, which paradoxically causes those same audiences to disrespect the press for pandering.

Of course, it does not help that the term “news” is frequently applied not only to fact-based reporting on newsworthy events but also to commentary, infotainment, and clickbait. Nor does it help that the purveyors of all of these forms of content often adopt the label of “journalists.” Even legacy media themselves have contributed to blurring the lines between these categories, making it unsurprising that the public mistrusts their motives. They have also contributed to the problem by, at times, constructing their own “filter bubbles,” hobbling their ability to connect with those they cover. In the meantime, demagogues stoke the public’s disdain for the press for their own political gain, portraying the press as biased and untrustworthy and encouraging partisans to brand any news with which they disagree as “fake.”Footnote ²⁷

8.2 Is Libel Law to Blame?

These factors have contributed to media distrust, but certain prominent voices point to an unexpected contributor to the eroding public trust in US media: American libel law exceptionalism.

In an era when the press faces unprecedented challenges, from financial instability to declining public trust, those who believe in the importance of a free press find themselves playing defense. Press advocates have been forced to articulate, with renewed clarity and urgency, why the press is not merely important but vital to an effective democracy and why it needs to be saved from what has been described as a “death spiral”Footnote ¹ and an “extinction-level event.”Footnote ² These arguments often focus on the crucial role the press plays – the “press function,” as lawyers call it – in informing the public, holding power to account, and facilitating democratic discourse. Advocating for the press’s value has become an essential task because the survival of this indispensable institution depends on the public understanding and appreciating its necessity.

Scholars and commentators generally discuss the press function through one of two lenses. Legal scholars tend to describe the press function through existing Supreme Court doctrine in which the Court has told us that the press serves as a watchdog, an educator, and a proxy for the public. Political theorists, meanwhile, will situate the press within our overall democratic structure, explaining that a free press is integral to democracy because it maintains the public sphere and facilitates discourse.

These explanations are fundamental. But they are also insufficient, especially in an age when delegitimizing the press is a political tactic. To deliver the press from extinction, the public needs not only to know what the press does, it needs to care.

This means that beyond conceptualizing the press function as a matter of doctrine and theory, press advocates must conceptualize it as a matter of rhetoric. Advocates should be rethinking how to describe what the press does when they are speaking to judges, legislators, and other citizens. These audiences can ensure – through enhanced First Amendment rights, public subsidies, public-media models, statutory protections, or subscription dollars – that the free press survives and flourishes instead of succumbs.

The act of naming is not a small thing. New language can persuade and motivate. Rethinking how we talk about the press – and press functions, in particular – could prompt Americans to care about the press and to find common ground regarding the reasons for this caring. Talking about the press in newly positive ways is also how the public might begin to envision a reinvigorated and dynamic institution. As communication scholar Mike Ananny writes, “[P]ress freedom is a normative and institutional product of any given era: it is what people think press freedom should mean and how people have arranged people and power to achieve that vision.”Footnote ³ What the “people think press freedom should mean” is shaped by the language they use. As of now, that language often does not account for the true value of the press’s work. This needs to change.

This chapter outlines how press advocates can think rhetorically about press functions in the service of a reinvigorated press and democracy. First, it proposes a shift in terminology from the dry “press function” to a descriptor that is not just more inspirational but more apt: “press benefits.” Second, it goes a level deeper and examines two time-honored press benefits – the watchdog and proxy roles – and suggests how these benefits might be renamed and reconceptualized. Third, it describes how we might use rhetoric as a framework for naming other press functions, even aspirational ones, that the courts have not previously recognized. This framework can be a means of both supporting the existing press and transforming it into one that is ever closer to best supporting the collective flourishing of citizens. Finally, the chapter ends by sketching out methods for both dispersing updated press-benefits rhetoric around the world and continuing to generate that rhetoric.

9.1 From Press Functions to Press Benefits

In naming what the press does, the Supreme Court has settled on the word “function.”Footnote ⁴ Sometimes the justices have made the reference a bit more grandiose by rendering it plural: “functions.”Footnote ⁵ Academics (including me) have followed suit and adopted the term.Footnote ⁶ The noun is accurate but uninspired. “Functions” are roles or purposes.Footnote ⁷ They are not inherently good or bad. They have no moral valence. Functions are anodyne, workmanlike, serviceable.

The press requires better. The work of the American press is not a mere collection of humdrum or trivial tasks. The press provides us with feats, public goods, and achievements, like documenting catastrophic flooding, revealing the scope of migrant child labor, explaining how dementia progresses and impacts families, and humanizing prisoners.Footnote ⁸

In doing this work, the press performs more than a function. It provides a benefit. A benefit is an “advantage, profit, good,” which, in the case of the press, redounds widely.Footnote ⁹ “Benefit” seems an especially fitting word given that journalism’s calling is to assist with the perpetuation of democracy. As Bill Kovach and Tom Rosenstiel write in The Elements of Journalism, “The primary purpose of journalism is to provide citizens with the information they need to be free and self-governing.”Footnote ¹⁰

This is not institutional self-aggrandizement. Countless political scientists, philosophers, and judges have described the tight tie between the press and government.Footnote ¹¹ The Supreme Court, in bygone eras, understood this link and the press’s vital role in sustaining democracy well. Take Justice Felix Frankfurter, who wrote: “Without a free press there can be no free society.”Footnote ¹² Or Justice Hugo Black’s Pentagon Papers concurrence that stated: “In the First Amendment, the Founding Fathers gave the free press the protection it must have to fulfill its essential role in our democracy.”Footnote ¹³ Given this acknowledged importance of the press to democracy, it seems right to refer to its role as a benefit.Footnote ¹⁴

By embracing the term “benefit,” press advocates would be using rhetoric in all three of the forms that law and literature scholar James Boyd White described as comprising “constitutive rhetoric.”Footnote ¹⁵ These include speaking the language of the audience, modifying and rearranging the language of the law, and creating a community around legal language.Footnote ¹⁶ In using rhetoric in all of these forms, advocates can spark doctrinal and cultural change.

With respect to speaking the language of the audience and rearranging the language of the law, “benefit” has an emotional appeal that “function” lacks. This emotional resonance is fundamental to convincing Americans to care. Though lawyers often resist acknowledging the salience of emotion in advocacy, it is the appeal to pathos that rhetoricians from Aristotle forward have recognized as a key to persuasion.Footnote ¹⁷ As one former New York Times op-ed page editor put it: When it comes to persuasion, “[f]eelings are crucial, much more important than facts.”Footnote ¹⁸ “Benefit” appeals to humans’ deep-seated desire for that which is good. Making this change is a way of appealing to the language of the audience.

Using “benefit” is also an example of constitutive rhetoric’s creation of community. The idea that the press provides benefits in service of democracy could help establish common ground among Americans who might otherwise remain polarized.Footnote ¹⁹ It is what legal scholar Joseph W. Singer calls a “public reason.”Footnote ²⁰ In the context of persuasion, public reasons are those ideas “that could or should be accepted by people with differing perspectives, religious traditions, moral frameworks, and experiences.”Footnote ²¹ If Americans can agree that the press is an engine of democracy, there is a basis for the “shared factual assumptions” and “shared values” that White describes as drivers of a rhetorical community.Footnote ²² If we can assemble this community, we have exponentially improved the chances of protecting and promoting the press.

It is this degree of sentiment and synergy that could help push through the inertia surrounding legislative efforts and legal strategies to protect the press. In Congress, press advocates have repeatedly fallen short of passing press-protective legislation.Footnote ²³ Meanwhile, in the Supreme Court, this inertia is manifest in a sort of Press Clause paralysis – a refusal by the Court to read any meaning into the First Amendment’s explicit reference to the press – and a more recent failure by the justices to speak positively about the press at all.Footnote ²⁴

At its root, the idea is that language evokes caring – a creation and realization of shared ground and community and ultimately an action in service of the thing (here, the press) about which the community cares. This action should plow back into improving the press, making it that much easier to care about. A shift in language can jumpstart this cycle and then fuel it.

9.2 Renaming and Rethinking the Watchdog and Proxy Benefits

Beyond the adoption of “benefits,” a relabeling and rethinking of some of those benefits themselves is in order. This effort is vital not simply so that we can describe the press more inspirationally but because it is long overdue. The Supreme Court opinions that set out the key press benefits are now about a half-century old – penned at a time when newsmen pounded out stories on typewriters in newsrooms ripe with the smell of stale cigarette smoke.

We can start with the watchdog. If there is a press function that has been most lauded by the Court and journalists themselves, it is the watchdog.Footnote ²⁵ It is the concept that the press’s primary purpose is a structural one – it exists as a check on the other three branches of governmentFootnote ²⁶ as well as a check on private power.Footnote ²⁷ Its primary role is to expose and attack government actors for bad behavior.Footnote ²⁸

This structural role is important, even necessary. But the watchdog label is flawed, and the prevalent belief that to be a watchdog is the press’s raison d’être is also problematic. The watchdog metaphor is an unnecessarily violent one, especially for an entity providing a public good and service.Footnote ²⁹ Watchdogs exist to sniff out wrongdoing and to pounce.Footnote ³⁰

The violent overtones of the metaphor are also evident in the way journalists routinely describe their role as hard-hitting and aggressive.Footnote ³¹ Watchdog reporting has been dubbed by some scholars as the “journalism of outrage” – designed to “prick[]” and “provoke.”Footnote ³² Without it, these scholars argue, journalism would be “sterile” and “bloodless.”Footnote ³³ Given the way that metaphor functions, there is a chance that clinging to this violent metaphor is actually preventing journalists and citizens from conceiving of the press in ways that are not so pugilistic.

Instead of describing the benefit that the press provides when it checks government wrongdoing as “watchdogging,” an alternative is to describe it as “independence.” This word eliminates the violent overtones of the watchdog yet preserves the idea that the press is acting, as New York Times publisher Adolph S. Ochs wrote in 1896, “without fear or favor.”Footnote ³⁴ An independent press can be both a monitor and a collaborator as situations demand. With its patriotic overtones, “independence” also aligns well with a reconceptualizing of press functions as benefits in service of democracy.

A relabeling of the watchdog function might also help lead to deeper, and overdue, thinking about the parameters of this particular press benefit. For example, as it stands, journalists conceive of the watchdog role in much the same way that First Amendment theory has traditionally thought about free speech – that more is better and that its virtues are almost without bounds. But just as recent changes in our political and speech environments have demonstrated that sometimes too much speech may have costs (see, for example, the damage done by “troll armies,” “‘flooding’ tactics that distort or drown out other speech,” or the “deployment of propaganda robots”),Footnote ³⁵ it is also worth examining whether there should be limits on the press’s watchdog role. For example, one recent study conducted by researchers from the American Press Institute and the University of Chicago found that conservatives may be less likely to embrace the press’s watchdog role.Footnote ³⁶ If this is right, the press’s emphasis on it may contribute to the public feeling alienated by the press and, as a result, failing to support it.Footnote ³⁷

Framing this inquiry in terms of “independence” suggests that there may be times and ways in which press independence may need to vary depending on the circumstances.Footnote ³⁸ For example, historically, journalists have staunchly opposed any sort of public funding precisely because of its potential to erode the press’s independence. Yet, at a moment like the present, when the press’s very existence might depend on public funding, press advocates are examining ways to publicly fund the press while ensuring that journalists retain editorial independence. In other words, “independence” can be nuanced. It is not all or nothing.

In addition to the watchdog role, another press benefit that could be relabeled and, in the process, rethought is the role of the press as a proxy. The term “proxy” has been used to refer to the press’s role as a literal stand-in for the public, for example, when journalists witness newsworthy events and then report on them.Footnote ³⁹ As the Supreme Court has written, “[I]n a society in which each individual has but limited time and resources with which to observe at first hand the operations of his government, he relies necessarily upon the press to bring to him in convenient form the facts of those operations.”Footnote ⁴⁰ Journalists have also long indicated that their proxy role involves gathering information from far-flung locales and bringing it to the public’s doorstep. As journalist Walter Lippmann wrote in 1922, “[T]he real environment is altogether too big, too complex, and too fleeting for direct acquaintance.”Footnote ⁴¹ Instead, he wrote, we “must have maps.”Footnote ⁴² The press, in its proxy role, creates those maps.

But, like “function,” the “proxy” label is uninspiring. Plus, it is perhaps not as accurate as it once was. Although there are still situations in which physical access to spaces is limited, the public now often has virtual access to newsworthy events. Meanwhile, today, serving as a proxy for the public on the world stage is an increasingly impossible feat given the rate of global information generation and sharing. Thus, what the public needs from today’s press is to act not so much as a proxy than as a “curator.”

By being a curator, the press surveys, selects, distills, prioritizes, and contextualizes.Footnote ⁴³ It does this work by adhering to the evolving tenets of journalism. This process stands in stark contrast – and its value becomes especially clear – when compared to the most dominant method by which the information that so many of us ingest today is sifted and presented: content moderation by social media platform algorithms.

To be clear, I am not suggesting that the watchdog or proxy roles are unimportant. But when we rely on these metaphors to describe the press’s work, we limit the public’s conception of that work, at times in unproductive ways.Footnote ⁴⁴ Using new language to describe the press’s benefits could broaden our understanding of the press’s potential and, in time, even transform the institution to fulfill that potential.

9.3 Press Rhetoric and the Public Imagination

So far, this chapter has discussed renaming press functions – that is, finding new ways to label press functions that the law has already recognized with different descriptors. But there may be additional, and even more powerful, ways to use rhetoric to bolster the press. We could employ rhetoric as a framework for generating language to describe press benefits that the law has failed to recognize or has not yet even imagined. In fact, this type of communal envisioning may be the very essence of conceptualizing law as a branch of rhetoric.Footnote ⁴⁵ As James Boyd White wrote, “It is the true nature of law to constitute a ‘we’ and to establish a conversation by which that ‘we’ can determine what our ‘wants’ are and should be.”Footnote ⁴⁶

Now is the right moment to be engaged in this process. Along with institutions generally, the press is neck-deep in crisis. It lacks the public’s trust and support.Footnote ⁴⁷ This may be for good reason. Communication scholars Barbie Zelizer, Pablo Boczkowski, and C. W. Anderson have critiqued today’s press as elitist, entrenched, and self-congratulatory.Footnote ⁴⁸ They write, “Journalism finds itself less socially, politically and culturally relevant than it ought to be, and certainly less relevant than it thinks it is.”Footnote ⁴⁹

But this crisis could be reframed as an opportunity for reimagining the press, starting with press benefits. Yes, as discussed, the press is a bulwark of democracy, but in what ways? How does the press inform, create communities, and empower? To borrow Mike Ananny’s wording, are the press’s “institutional arrangements” such that they will “produce expansive, dynamic, diverse publics”?Footnote ⁵⁰ If not, how might we change them? What publics do we want, and how do they differ from what we have? How might the press better help with solving collective problems? How might the press better connect with and serve various communities? Is the press promoting communal flourishing? What type of democracy is our press a bulwark for? Is this the type of democracy we want and need?

In trying to answer these questions, we might both discover and imagine all sorts of ways that the press is benefiting citizens or, conversely, is not benefiting them and needs to be. For example, we could focus on the role of the press as a facilitator and convener of the public square. In this role, the press creates the shared epistemic foundations that are necessary for democracy.Footnote ⁵¹ We might also recognize the press’s role as a historian, creating vital records through which future generations will be able to study and build upon our generation’s thoughts, advances, and missteps. We could imagine the press as a community storyteller, a preserver of traditions. We could envision the press as a sentinel and observer whose very presence improves outcomes. Perhaps we see the press as creating a kind of poetry – language grounded in fact but allowing us to see possibility.Footnote ⁵² We might aspire to a press that tends to communities through care-based practices.Footnote ⁵³ Or we might decide these roles are too big or too many for the press, and we might think about how to distribute them or what other institutions or organizations or arrangements might fulfill them.

Even in the short term, this use of rhetoric and these conversations could help to create the kind of dynamic institution that some press thinkers argue is a productive path forward. Harnessing the public imagination is key. According to communication scholar Stephen D. Reese, “institutions are worth defending to the extent they contribute to a moral purpose and bind together historically the aspirational values they embody, connecting members to their society – past, present, and future.”Footnote ⁵⁴ As part of creating that worthy entity, Reese argues that we should think about the press as a “hybrid institution” – an entity that blends old and new and is constantly engaging, evolving, and emerging.Footnote ⁵⁵

The key to Reese’s hybrid institution and to employing rhetoric more generally is not doing so with the goal of arriving at definitive answers to the questions posed. More important is the dialogue – the community it creates and the changing institution that the community can sustain. This is, at its core, the messy work of democracy. As philosopher Jason Stanley writes, democratic citizenship “demands a great deal of all of us. There are easier ways to live.”Footnote ⁵⁶ It may be that these conversations about the press prompt citizens to think about democracy in new ways. Our current press is a product of our particular American brand of liberal democracy, but different types of democratic government might demand other press benefits, roles, and forms.Footnote ⁵⁷ And regardless of what shape our future democracy takes and what the needs of that democracy might be, building rhetorical frameworks to support the press both normatively and practically is a crucial task.

9.4 Where to Use Pro-Press Rhetoric

As a new language around press functions is generated, the beauty of this rhetoric is that it can be deployed without cost and in abundance. It can take numerous forms and be directed at a variety of audiences.

Let’s start with the judiciary. Given the judiciary’s ability to shape the parameters of the press and invest it with power, it is critical that judges understand the value of the press’s work. One direct way to demonstrate this value is through briefing. Press advocates can advance their cause by adopting updated language about the press in their merits briefs and in the numerous amicus briefs they file. As constitutional law scholars Eric Ruben and Joseph Blocher write, rhetoric is at the heart of legally instigated change: “Constitutional doctrine is shaped by the rhetorical moves that litigants, commentators, scholars, and judges make. Such rhetoric can become embedded in popular and legal understandings of constitutional rights.”Footnote ⁵⁸

Advocates might also speak this language to legislators. Both at the federal and state levels, legislators convene hearings on the state of the press as well as draft and debate press-related legislation. Pro-press rhetoric in testimony and advocacy can find its way into law and policy.

Journalists, as public figures, must also become more accustomed to using their megaphones to bring attention to press benefits. Institutionally, the press has a bias against itself being at the center of the news.Footnote ⁵⁹ Other than having a byline, some journalists want to disappear behind their stories. Traditionally, this is part of the feint of “objectivity” – or the so-called “view from nowhere.”Footnote ⁶⁰ But as the premise of objectivity has come under fire, perhaps, too, should the norm among journalists to keep relatively mum about the fate of their profession. As one journalism professor recently wrote in the Columbia Journalism Review about convincing conservative audiences of the press’s value, “If local journalism is really to be revived, those who hope to revive it will need to try something more ambitious: they’ll need to interject in the conversations about journalism within conservative networks, and offer a compelling new explanation of who local journalists are and why conservatives might want to listen to them.”Footnote ⁶¹

Employing pro-press rhetoric in these public ways and spaces is crucial both as a means of dispersing it and of continuing to germinate conversation about the press. In addition, it can perhaps blunt the effects of the rabidly anti-press rhetoric that has become a hallmark of numerous conservative politicians. To date, no one has effectively countered the populist rhetorical tirade against the press. To do so, we need both a reinvigorated language around the press and people willing to speak it.

9.5 Conclusion

In many ways, the American public has forsaken the press. This is obvious from vitriol – like the January 6 rioter who wrote “MURDER THE MEDIA” on a door of the U.S. Capitol.Footnote ⁶² But the more pernicious acts have been banal acts of omission, such as failing to provide financial support to the thousands of local news outlets that have gone dark in the past two decades.Footnote ⁶³

This unintentional forsaking is perhaps understandable. The press survived, and profited, for most of America’s history without significant public intervention. It amassed power through the wealth it reaped from advertising. For the most part, it has not demanded much of citizens. But that heyday has passed. If Americans want a functioning press (which is to say, if Americans want a functioning democracy), we have to learn to do the work of supporting it.

One way we can all contribute is with rhetoric. As philosopher and novelist Iris Murdoch wrote, “[L]anguage itself is a moral medium, almost all uses of language convey value.”Footnote ⁶⁴ Ensuring we have a free press has significant moral stakes; how we talk about the press needs to convey its value. If we can conjure the language that describes the press we need and want, we come closer to bringing that press into being.

As I write this in May 2024, there is, yet again, anxiety about the power that an emerging technology, namely generative artificial intelligence (or GenAI), has to shape journalism, news, and the viability of a robust, economically healthy, and free press. While earlier panics focused on the influence of the internet and digital data on journalistic practice,Footnote ¹ social media’s control of attention and advertising,Footnote ² and the power of complex and proprietary algorithms to create “echo chambers” and “filter bubbles”Footnote ³ that undermine journalism’s role in democracy, the contemporary moment features a potentially more existentially destabilizing force.

Depending on which technologist, activist, scholar, or pundit you read,Footnote ⁴ the prediction of how GenAI is likely to influence journalism varies. It may free journalists from mundane tasks and empower them to pursue more complex reporting, or it may further shrink an already atrophying labor force struggling for stable work. It may infuse internet search results with high-quality news that has previously been scraped or licensed from publishers, or it may sequester news into opaque datasets and probabilistic language models that deliver neither reliably true results nor traffic to news advertisers. GenAI’s propensity to haphazardly extract, combine, and hallucinate data may even further erode online public spheres, or it may privilege publishers as producers of high-quality news created in the public interest over the generic “big data” that technology companies greedily need for general computational models and advertising markets.

In many ways, GenAI is not a new threat to skilled workers like journalists. In his 1978 study of how workers responded to automatically controlled machine tools, David Noble identified three ways that technological innovations challenge workforces,Footnote ⁵ all of which apply to journalists in this moment of GenAI. First, workforces must “transcend the ideology of technological determinism by demystifying the technology itself.”Footnote ⁶ Journalists often critique and attempt to undercut the fantastical claims that some technologists – and some fellow journalists – make about GenAI’s power to produce believable, useful, and democratically forceful language.Footnote ⁷ Second, the workforce must “regain its confidence by preparing itself” and articulating “its own choices in its own … interest.”Footnote ⁸ Journalists might differentiate what they do – reporting, writing, and editing with sophistication and in the public interest – from what GenAI, which is inscrutable and incapable of sentience, professionalism, or reflection, does, namely mimicking the statistical patterns of datasets and models that produce media. Finally, workers must “struggle, on the shop floor and politically, to get into the position to make these choices and thereby steer the course of ‘progress.’”Footnote ⁹ That is, journalists must not only debunk the fantastical claims of GenAI proponents and differentiate their work from GenAI’s outputs but must also be able to act in their own interests and know what those interests mean in different technological contexts. They must be able to practice their craft in ways that reflect what they think their profession and public obligations demand. This capacity to act with knowledge and intent requires, as a preliminary condition, knowing well enough how a technology works to appreciate how it matters to professional practices. Only then can professionals appreciate what they should do with a technology and make informed and intentional choices about resisting or reshaping it and/or their professional standards.

This capacity to act with knowledge and purpose is a kind of press freedom. An autonomous press knows its systems well enough to reshape them in its own professional image, or to refuse them, and also understands its public obligation. A press that is beholden to a technology’s fantastical claims or moral panics, or that cannot separate its own vision from that of the technologists, will be unable to structure and change itself, a hallmark of any profession’s autonomy. By “autonomy,” then, I mean the profession’s capacity to make its own decisions about its practices – for example, what to cover, how to write, who to interview, when to publish, or which ledes to foreground. An autonomous press can see its own actions, reflect on them, and change them when it chooses to.

To explore how well the press can change itself, I focus here on how the press responds to mistakes by examining moments when the press has investigated, corrected, suffered, or tolerated some failed or broken-down aspect of either its own journalistic work or the larger information environment it operates within. My contention is that a strong, autonomous, self-empowered, and self-corrective press has the capacity to report and publish on mistakes – its own and others’ – and that GenAI poses a particular threat to this capacity because so many aspects of synthetic media and its use in journalistic work are currently unknown.

Taking seriously the idea that an autonomous press self-monitors and self-corrects, we might question whether the institution has recursive press freedom – a capacity to create for itself the structures that enable journalism and serve the public interest. A recursively free press would confidently trace and shape the social, cultural, political, and technological forces that structure journalism and create news without needing to rely on outsiders for knowledge, permission, or capacity. Such recursive freedom would be especially critical in the context of the synthetic press, which refers to the journalistic practices, forms, and institutional conditions that are intertwined with GenAI and often unknowable and unchangeable within proprietary datasets, machine learning models, and private corporations. Recursive, synthetic press freedom would be the press’s own capacity to shape GenAI according to its own journalistic mission and its own sense of public service.

This chapter investigates what this recursive, synthetic press freedom could look like in three parts. First, it briefly frames the current GenAI challenge to press freedom in terms of long-standing scholarship on journalism’s relationship to technological change. Second, it presents a case study to illustrate this issue, focusing on how the free press pursues and self-corrects its own mission and how it deals with “error” in a variety of ways. Third, the chapter uses instances of recent GenAI errors to demonstrate how this new technology hampers the press’s ability to know and change itself, thus damaging its autonomy. And, finally, inspired by Noble’s study of how workers might regain autonomy in the face of outside technological change, the chapter sketches a model of “recursive press freedom” that all defenders of journalistic autonomy might adopt to understand and reshape this GenAI moment.

I hope that different scholarly audiences will find this chapter useful in multiple ways. Journalism scholars, for example, are rapidly endeavoring to understand the relevance and usefulness of GenAI to news production and its impact on normative ideals of the press.Footnote ¹⁰ Examining instances when GenAI has failed gives us a way to think about what a successful press is presumed to do and what good or proper uses of GenAI in journalism might be. When scholars or practitioners foreground journalistic errors and call for remedies, they reveal their capacity to monitor, correct, and envision a better press. The press’s power to self-correct – to define, prioritize, and fix its own errors – is a key aspect of press freedom, and one that is increasingly critical as many aspects of reporting, editing, and publishing become intertwined with technological infrastructures existing beyond the control of newsrooms.

This is an important moment to develop a precise understanding of how technologies influence journalism because there are ramifications for understanding the scope and meaning of press rights. As Frederick Schauer cautioned, failing to appreciate the “sociological messiness of institutional demarcation”Footnote ¹¹ – for example, how blogging, investigative journalism, pornography, sports broadcasting, or recommendation algorithms all pose different speech challenges – may cause a First Amendment “institutional ‘compression’”Footnote ¹² that leaves everyone with fewer rights than they might have had if journalists, technologists, and courts alike had better understood the significance of drawing the press’s boundaries and of defining “the press” narrowly as what journalists do or more broadly as what technologies, technologists, and online audiences also do. Focusing on journalism’s disposition toward errors and mistakes is one way of achieving empirical precision and of understanding how different institutions and forces constituting the press collide. This is a challenge ideally suited for scholars of science and technology studies, critical internet studies, and platforms studies who have long traced how technological practices, actors, values, and politics intersect to create newsrooms, professional journalism, audience cultures, and news economics.Footnote ¹³ A focus on journalistic errors and GenAI may reveal how different social and technological actors define mistakes, debate causes, classify harms, anticipate risks, assign responsibility, and apply remedies. If press freedom emerges not only from intersections between journalism and jurisprudence but also among messy collisions – of people, professions, norms, institutions, data, and computation – then analyzing mistaken meetings and failed alignments should illustrate what press freedom could be if the power and politics of digital media were otherwise.

10.1 Press Freedom as an Institutional, Infrastructural Achievement

To better understand what press freedom means in this era of GenAI, it is essential to first understand how the “the press” is always a product of an era’s institutional and technological forces and how the concept of “press freedom” always depends on ideals relating to its public power and obligations.

Though it is beyond the scope of this chapter to review the large, long-standing, and interdisciplinary research tracing the various forces that have constituted “the press,” it is worth highlighting how three dimensions of the press – news forms, journalistic practices, and institutional relationships – have emerged from the collisions of social and technological forces, with this providing an expansive and flexible view of the press that may help stave off the compression of rights that Schauer cautioned against.

First, news forms – text, image, sound, video, virtual reality, games, and participatory forums – have always been intertwined with advances in media representations and publishing technologies. Advances in printing, layout, and graphics transmission enabled newspapers to develop brand-specific visual designs, direct readers’ attention, sustain audience engagement across a newspaper, and use high-quality imagery to signal both journalists’ authority to report and the authenticity of their accounts.Footnote ¹⁴ Subsequent advances in synchronized audio and visual content, chyron overlays, animated graphics, and data visualizations allowed news organizations to create immersive, layered news experiences that could display and interweave different information streams simultaneously.Footnote ¹⁵ Developments in transportation infrastructure, satellite networks, and lightweight cellular technologies enabled “live news” to emerge as a unique genre. And the popularization of communication technologies amongst readers – encompassing everything from cheap and reliable postal service and home telephones to online bulletin boards and social media platforms – spurred the creation of letters to the editor, call-in shows, and comment threads that blurred the distinction between the journalistic voice and public participation.

Every new technological advancement has prompted the question of what it means for news stories to be authoritative, authentic, timely, persuasive, and professionally produced. Put differently, to the extent that press freedom depends upon and enables journalists’ ability to create stories that are defensible as forms of information – for example, that are factual, free of mistakes, and not manipulative – that freedom is always inseparable from an era’s communication technologies and media representations.

Second, journalistic practices – fact-finding, interviewing, witnessing, writing, editing, and publishing – have always reflected how an era’s communication technologies influence epistemological standards, modes of working, and professional norms. Publishers in the colonial era invested in boat-building techniques to produce small ships that could quickly reach ships arriving from England with news, mail, and commodity prices, ensuring that their newspapers had the timeliest information and feeding early anxieties about being scooped by competitors.Footnote ¹⁶ Advances in literacy, printing, distribution, industrial advertising, and market segmentation all fed the emergence of journalistic objectivity, not as an ideal of impartiality and public service but as an economic practice that let publishers avoid offending any one set of readers and, most importantly, advertisers.Footnote ¹⁷

As technology has continued to advance, so too have modern debates about these new journalistic practices. Computer-assisted reporting and data-focused journalism have allowed investigative reporters to portray themselves as disinterested scientists reliably and defensibly discovering patterns of injusticeFootnote ¹⁸ consistent with the moral standards of their times.Footnote ¹⁹ And now that online publishers can track how well their stories are performing in search engine results, digital advertising markets, and on social media platforms, newsrooms must grapple with how to do journalism that is aware of, but not beholden to, internet traffic patterns.Footnote ²⁰ Similarly, journalists are wrestling with whether social media platforms like XFootnote ²¹ are places to seek out stories, find sources, speculate, publish independently, develop personal followings, or some combination thereof.Footnote ²² Advances in drone technologies, meanwhile, have opened new events and regions for journalists to observe, but they have also invited skepticism about the value of such high-altitude witnessing versus grounded reporting.Footnote ²³ Encrypted tools like Signal, PGP keys, and the Tor network allow journalists to connect securely with potential sources and whistleblowers, and while the process is still cumbersome,Footnote ²⁴ increasing efficiencies in Freedom of Information Act (FOIA) requests give journalists more access to public information more quickly, enabling new forms of public oversight reporting.Footnote ²⁵

Fast boats, mail service, computer databases, search engine analytics, social media platforms, drones, and encryption are just a few of the myriad technological advancements that have influenced journalistic practice, and it would be a mistake to see them as separate from how journalists make judgments about newsworthiness, facticity, audience engagement, public service, or source ethics. If press freedom is, in part, a journalist’s ability to decide the workflows, norms, and practices that they believe align with their professional judgment, then it is impossible to separate these decisions from the communication systems that suggest actions, encode standards, enable practices, and generally structure information work. Is it a mistake or an infringement of press freedom if a journalist subconsciously trusts a source with a PGP key, a Twitter user with a blue checkmark, or an organization highly ranked by a search engine? These seemingly trivial examples point to a messy and implicitly structured space of journalistic practice that is inseparable from communication technologies.

Finally, institutional relationships – the press’s connections to governments, markets, industries, and audiences – have always influenced how the press has understood its mission and scoped its autonomy. Defining institutions as “social patterns of behavior identifiable across the organizations that are generally seen within a society to preside over a particular social sphere,”Footnote ²⁶ Cook showed how “the press” is actually a set of relationships among not only news organizations but also government officials, public relations professionals, civil society actors, political parties, and other communicators interested in structuring communicative self-governance. These relationships, according to Cook, need and sustain the institutional practices of mimicry, isomorphism, and labor division to distribute the work and power of the press across largely invisible backstage processes.Footnote ²⁷ Instead of seeing the press as an autonomous institution living within news organizations, professional bodies, and public ideals, we might more correctly see it as a pattern of distributed, intertwined communication practices that seem coherent to the extent that they resemble what a given era socially, culturally, economically, and technologically expects “the press” to be. This coherence is stabilized when, for example:

• a government spokesperson frames events and gives quotes in ways that resonate with a journalist’s expectations and a publisher’s conception of its audience;Footnote ²⁸

• public relations representatives position press releases and product launches to fit not only with their clients’ goals but also with journalists’ understanding of industry novelty and technological innovation;Footnote ²⁹

• news organizations perennially realign their judgments of newsworthiness and public service with social media platforms’ own assessments of audience engagement and advertiser value (as seen in publishers’ “pivot to video” to meet Facebook’s expectations and contemporary investment in short videos that perform well on TikTok and YouTube);Footnote ³⁰

• fact-checking organizations structure their rhythms and standards of verification to align with the goals of the social media platforms that spread disinformation, creating a symbiotic service arrangement in which fact-checkers are essentially platform partners;Footnote ³¹ and

• publishers like the Associated Press and NewsCorp license their content to OpenAI, ensuring that future ChatGPT results will somehow include or account for the journalism produced by those companies.Footnote ³²

These examples of the institutional perspective on press freedom may seem like reasonable arrangements that simply ensure journalism’s survival and relevance in rapidly changing media systems. Without such alignments, patterns, and partnerships, the press might simply disappear. But they also suggest that a common-sense understanding of “the press” as comprising what journalists do and what publishers distribute is simply incorrect. The press is actually a set of largely invisible, strategic relationships with their own logics of alignment, their own standards of quality, and their own senses of what it means to tell factual stories in the public interest.

This image of “the press” as a powerful and messy interplay of social and technological arrangements has gained traction in recent years. Early digital newsrooms were shaped by mixes of journalistic and technological mindsets and practices.Footnote ³³ Digital algorithms quickly influenced reporting and publishing practices.Footnote ³⁴ As social media rose in prominence, a hybrid media system of seemingly disparate communicators emerged, which journalists both participated in and reported on.Footnote ³⁵ Histories of the press show how its last fifty years have been inseparable from the datasets and computational epistemologies that journalists use to know their world.Footnote ³⁶

My own work has examined how the press’s digital infrastructures enable the public to hear the perspectives of others and receive information that they would likely not seek out for themselves,Footnote ³⁷ and Kate Crawford and I have suggested that the press is a “liminal” achievement somewhere between journalistic judgment and technological design.Footnote ³⁸ Emily Bell and Taylor Owen have identified a “platform press” that has emerged from publishers’ seemingly endless responses to new technologies and policies.Footnote ³⁹ Matt Carlson and Seth Lewis have traced the contemporary press through the complex boundary work of diverse laborers,Footnote ⁴⁰ which Adrienne Russell has shown also includes digitally savvy activists, environmentalists, and solution-oriented journalists.Footnote ⁴¹ Stephen Reese has demonstrated how the crisis of the institutional press is a problem not only of journalism and publishing but also of information economics and political culture.Footnote ⁴² Newer work centered on artificial intelligence is revisiting these questions, with a focus on how computation, technologists, datasets, and experimental methods are currently structuring the press, its language, and its own understanding of its public service role.Footnote ⁴³

It is impossible in this chapter to fully capture the extent of this literature, but there is ample scholarship indicating that “the press” is and always has been a messy and historically situated mix of colliding social and technological forces. The idea, then, that “press freedom” is only, or even mostly, a matter of journalism and jurisprudence is incorrect. It is more accurately understood as an ongoing achievement of multiple, intertwined actors, existing in historical-technological moments, which create separations and dependencies that are continually judged against normative ideals of what journalistic autonomy is supposed to be and what journalism’s public mission should be in any era.Footnote ⁴⁴

If these conceptions of the press and its freedom form a plausible framework, then the challenge becomes not only how to empirically observe these collisions of forces – these separations and dependencies – but also how to judge whether they move us closer to or further from any particular ideal of press freedom.

I want to bracket the discussion of which version of “press freedom” is worth pursuing amidst this mix of forces – much excellent scholarship has analyzed the term and the significance of various definitionsFootnote ⁴⁵ – and instead argue for a way of seeing press freedom, namely, as the press’s capacity to self-monitor and self-correct and to learn from its mistakes. Taking inspiration from Samuelson’s call for a “freedom to tinker,”Footnote ⁴⁶ scholarship on how “broken world thinking” offers new ways to see system maintenance and repair,Footnote ⁴⁷ studies of institutional learning,Footnote ⁴⁸ and organizational technologies,Footnote ⁴⁹ I wish to suggest that a largely ignored aspect of press freedom is the press’s ability to learn from mistakes. While some works have encouraged journalists to experiment with GenAI, learn how it can be useful for news work, and avoid making mistakes while using it,Footnote ⁵⁰ I want to invoke a larger, more generative sense of experimentation to ask how the press might learn about itself by critically engaging with how GenAI mistakes happen, why they matter, and what they might mean for press freedom in our contemporary era of synthetic media. If an autonomous press can self-monitor and self-correct – knowing its own forces and reshaping them when they fail – then the autonomy of the synthetic press (the journalistic practices, forms, and institutional conditions intertwined with GenAI) depends upon how well it knows, learns from, and changes in response to its mistaken engagements with GenAI.

The remainder of the chapter argues that the press has always defined itself through relationships with “mistakes” (broadly construed), that the contemporary synthetic press is struggling to understand itself through GenAI failure, and that a press that can learn from its mistakes has a kind of “recursive press freedom” that will be essential for defining its own normative vision amidst continued technological turmoil.

10.2 The Press and Mistakes: Its Own and Others’

In many ways, the press feeds on mistakes. Mistakes can be targets for correction, triggers for outrage, focuses of investigations, opportunities to teach audiences how knowledge develops, chances to develop and defend the profession, indicators of mistaken institutional relationships, and even matters to strategically ignore. While a complete review of how the press deals with mistakes, errors, and failures is beyond the scope of this chapter,Footnote ⁵¹ as context for understanding GenAI journalism mistakes, I offer five ways to understand the press’s relationship to error: as (1) misinformation to correct, (2) transgressions spurring outrage, (3) professional malpractice and self-regulation, (4) misconfigured publishing systems, and (5) tolerated ambiguities.

(A note on terminology: I use the word “mistake” somewhat loosely and interchangeably with “error” or “failure” or “breakdown.” This is not because I see these terms as synonymous – indeed, an extensive body of interdisciplinary literature differentiates these words more preciselyFootnote ⁵² – but because I mean for this chapter to broadly assert that the press’s freedom requires a capacity to self-correct and to structure and change itself in relation to things that it sees as somehow wrong, failing, broken, or misaligned with its vision of ideal professional journalism or healthy, democratic, communicative self-governance.)

10.3 The Synthetic Press and Its Mistakes

The synthetic press, which encompasses journalistic practices, news forms, institutional relationships, economic models, and regulatory frictions prompted by GenAI datasets, large language models, and commercial products, presents a new space of mistakes, errors, and failures for thinking about press freedom.

A considerable body of recent scholarship has traced how journalists report on GenAI,Footnote ⁸³ how newsrooms adopt and experiment with GenAI,Footnote ⁸⁴ how media organizations establish GenAI guidelines,Footnote ⁸⁵ how media guilds bargain around GenAI,Footnote ⁸⁶ and how GenAI poses new normative and regulatory challenges.Footnote ⁸⁷ But there has been little work so far framing GenAI as a press freedom problem.Footnote ⁸⁸ Thus, the focus of the present work is to understand how the press’s mistaken or failed uses of GenAI might repeat or break its historical patterns of contextualizing, containing, tolerating, or ignoring mistakes.

Beginning around the fall of 2022, GenAI began appearing in new, popularly available systems like ChatGPT, Midjourney, Copilot, DALL-E, Meta AI, Claude, and Sora. It was already at work in familiar tools like Google Search, Microsoft Office, and Google Docs and in countless, largely invisible, back-end integrations with existing organizational information systems. Journalists were quick to use and misuse the technology.

Indeed, examples of mistaken or failed GenAI journalism already abound. The German magazine Die Aktuelle fired its editor-in-chief after he published a synthetically generated “interview” with race car driver Michael Schumacher, which gave audiences no indication that it was a fabricated story.Footnote ⁸⁹ The Guardian discovered that ChatGPT was delivering plausible, but entirely fictional, news stories that the technology claimed the news organization had published.Footnote ⁹⁰ USA Today apologized and corrected countless sports stories after discovering that many of them were synthetically generated and contained errors.Footnote ⁹¹ CNET paused using GenAI after it published a series of stories offering readers financial guidance that contained myriad mistaken formulas and calculations.Footnote ⁹² Gizmodo found itself running afoul of Star Wars fan communities after it used GenAI to publish several stories with errors about the franchise’s characters and plots.Footnote ⁹³ Microsoft was harshly criticized after its GenAI news platform placed a poll next to a Guardian story, inviting readers to speculate whether the subject of the story died by murder, accident, or suicide.Footnote ⁹⁴ Sports Illustrated backtracked on its use of GenAI after an investigation found that it had published an undisclosed number of entirely fabricated stories under the bylines of synthetically generated journalist personas. The magazine blamed a third-party content provider for the mistake.Footnote ⁹⁵ Gannett paused its use of GenAI after it published several synthetically generated articles that contained information placeholders in place of story details (for example, running the lede “The Worthington Christian [[WINNING_TEAM_MASCOT]] defeated the Westerville North [[LOSING_TEAM_MASCOT]] 2-1 in an Ohio boys soccer game on Saturday.”).Footnote ⁹⁶ And MSN.com recently mistakenly featured a story falsely accusing an Irish broadcaster of sexual misconduct, a story that had been synthetically generated by a third-party chatbot that paraphrased real news stories.Footnote ⁹⁷

While some individual errors are amusing and idiosyncratic growing pains of technological experimentation, together they suggest that the synthetic press’s inclusion of GenAI yields new types of journalistic mistakes and thus new ways to think about press freedom through GenAI mistakes. Note, for example, how these mistakes span questions of editorial judgment, failed third-party relationships, factual errors, journalistic taste or decorum, specialized knowledge, reporter identity, and system design. These instances of mistaken GenAI journalism all deserve deeper theorization and case studies, but taken together, they suggest a framework for thinking about synthetic press freedom. In the following, I briefly sketch three sites of GenAI journalistic mistakes, suggesting each site as a place to consider how press autonomy might appear in synthetic media systems.

10.4 Journalistic GenAI Mistakes as Synthetic Press Freedom Dynamics

10.5 Conclusion: Recursive Press Freedom for the Synthetic Press

Through this chapter, I have developed three lines of argument. First, I suggested that “the press” is most correctly understood as a set of intertwined social, cultural, political, and technological forces. Although often primarily associated with the work of journalists and the production of news, the press is, in reality, an entity shaped by the actions and influences of multiple actors.

Second, I argued that a key element of press freedom is the press’s capacity to define, uncover, explain, control, fix, or ignore mistakes. Broadly understood, mistakes are a key means by which the press knows and regulates itself, as well as a lens through which it observes and endeavors to change society. Mistakes are an important and largely understudied way of understanding the press and its power.

Finally, I have tried to argue that, today, the press is a synthetic press to the extent that GenAI, through its datasets, models, and products, structures journalistic practices, news forms, institutional relationships, economic models, and regulatory frictions. And just as past eras of the press had their own types of errors, the synthetic press has its own emerging set of mistakes that manifest themselves in how journalists engage GenAI tools, how past errors echo in GenAI infrastructures, and how dataset politics structure journalism and news.

This conception of press freedom as the capacity to know, control, make, and learn from mistakes is not a stand-alone model of press freedom but is deeply connected to more general ideas about how institutions achieve their autonomy and how the public can govern themselves through communication technologies.

As Schauer argued,Footnote ¹²¹ institutions discover and demonstrate their independence, in part, through the capacity to create and control exceptions. When institutions reject default rules, carve out special cases, reverse precedents, and use their own judgment to evaluate the significance of their circumstances, they distinguish themselves from other institutions and demonstrate their ability to act with self-awareness, purpose, and confidence. A free press should see mistakes as exceptional opportunities to know itself better, to judge whether an error illustrates a pattern to tolerate or change, and sometimes to argue that a mistake is not its own but someone else’s failure to treat it differently. A free synthetic press, for example, would make its own GenAI infrastructure that fixes, prevents, or tolerates errors as it sees fit, and a free synthetic press would reject or work to change GenAI infrastructure that causes errors that GenAI may tolerate but journalism cannot. In practice, and aligned with calls for institutional exceptions to data-driven rules,Footnote ¹²² synthetic press autonomy would demand freedom from unauthorized data scraping, unethical datasets, deceptive interfaces, and anachronistic language models and freedom to create its own GenAI systems. A free synthetic press would create exceptions for itself from general-purpose GenAI because the mistakes that would otherwise result would be unacceptable.

As Chris Kelty contended, publics become “recursive publics,” when they have the “discursive and technical”Footnote ¹²³ means of creating and recreating themselves. Put differently, they can talk about who they are and why they are structured as they are, and they have the expertise and material power to sustain and change themselves without relying on outsiders. A simple example of a recursive public is an online forum composed of people who have the communicative capacity to talk about why and how they convene and who have the engineering expertise to maintain or redesign the forum, all without the need to consult or rely upon external actors.

A recursively free press, then, is a press that knows and can confidently discuss the social, cultural, political, and technological relationships that structure its journalism and produce its news and that has the power to change those relationships with intention and without needing outsiders. A recursively free press, for example, would not try to emulate Mark Zuckerberg’s model of community and would not need Google’s funding. A recursive synthetic press would deeply understand GenAI’s datasets, models, and products; would understand and be able to discuss their significance to journalism and news; and would be able to reject, change, or invent a GenAI infrastructure that it desires, without relying on outsiders. Such a press would have command of its future and be able to, as Erin Carroll has argued, “conjure the language that describes the press we need and want.”Footnote ¹²⁴

Artificial intelligence pioneer and ultimate critic Joseph Weizenbaum wrote that “the computer scientist has a heavy responsibility to make the fallibility and limitations of the systems he is capable of designing brilliantly clear.”Footnote ¹²⁵ The same might be said for the GenAI journalist, who should know how and why the synthetic press fails, which failures matter, and how it might be fixed. Such a journalist would hold an image of journalism, debate it with others, and trace it across datasets, models, interfaces, economics, politics, ideals of the public, and imaginings of what the synthetic press could be if it were autonomous.