Plus: Building trust through smiles, late-night humor encodes facts, and will the public pay for AI journalism?

Social media — for those of us old enough to remember, “Web 2.0” — was once hailed as a democratizing force.

The reality has been, of course, both contradictory and convoluted. It turns out that putting communication power in the hands of the people enables not just social justice movements like Black Lives Matter and the Arab Spring, but also a populism that’s decidedly authoritarian — a public clamoring for strong-man governance that, ironically, suppresses the populace.

A democratized information sphere not only allows populist activists and sympathizers to supply much of the world’s content, but also lets them act as “secondary gatekeepers” who help determine what mass media content everyone else sees (in part supplanting traditional media gatekeepers). The result has been populist uprisings such as the January 6, 2021, attack on the U.S. Capitol, the 2023 occupations in Brazil, and the election of authoritarian leaders on multiple continents.

Observing these concerning trends, five researchers decided to more deeply examine how support for right-wing authoritarianism relates to reasons for people’s news consumption and news sharing. Marcos Paulo da Silva, Rachel R. Mourão, Tim P. Vos, Marialina C. Antolini, and Gisele S. Neuls’s paper, “News characteristics, newsworthiness and secondary gatekeeping in Brazil: Influences of right-wing authoritarianism,” hit the Digital Journalism website last month. The authors surveyed 1,225 Brazilians one week after the 2023 Brasília riots, when participants violently occupied the National Congress, Presidential Palace, and Supreme Court buildings.

The researchers found that the more people supported right-wing authoritarianism, the more their willingness to consume news depended on three perceived qualities of the news. These were appearance (whether the headline looked professional and trustworthy), credibility (whether the news was from a credible outlet, was by a respected journalist, and was shared by a trusted acquaintance), and newsworthiness (whether the news captured the participant’s attention, was important to them, and aligned with their views). In contrast, right-wing authoritarians did not value confirmability — whether they could check that the news was true.

Turning to sharing, the study found that right-wing authoritarians based their sharing only on newsworthiness. The newsworthiness definition, in the sharing context, included whether the item was important for and could entertain other people. Authoritarians did not tend to share based on appearance, credibility, or confirmability. “Checking for correctness is irrelevant to them,” the authors concluded.

There are more intriguing findings we don’t have the space to explore in depth here. The authors investigated the relationship between “news repertoires” — people’s self-created bundles of regularly used news sources — and the four perceived aspects of news. And they found that frequent social media users were more likely to value all four aspects (appearance, confirmability, credibility, and newsworthiness), compared to less frequent users.

The authors say their paper shows not only the continued importance of secondary gatekeeping, but the varied ways in which this is practiced. “Support for right-wing authoritarianism relates more to the message’s perceived importance and alignment with personal views than to its truth or accuracy,” they write. This suggests a worrying feedback loop, as authoritarians may tend to spread misinformation that then pushes others to adopt their views and their movement. There is room for doubt, however, and for hope. The authors note that their method is correlational, and any causality is assumed rather than shown. The survey timing, just after the Brasilia riots, may have led participants to respond more emotionally.

But considering that misinformation often has the appearance of news without the confirmability, these results add to an important body of evidence. Namely, we’re finding that traditional gatekeeping norms have been repurposed and at times discarded by today’s secondary gatekeepers — i.e., everyone with a phone in their hand — and verifiable truth seems to be the first victim.

Research roundup

Newsrooms across the globe are increasingly turning to generative artificial intelligence to shape portions of their content. Just two years ago — an eternity in this fast-moving space — a survey found that more than half of newsrooms had already adopted AI tools such as ChatGPT. That share is certainly rising by the day, as AI is used for editing, for headlines, for content creation itself. We know much about the supply of this content — who produces it, how it is produced, even why it is produced. But we know far less about the *demand for this content *— how audiences perceive it. That gap is where this experiment-based study breaks ground.

The experiment drew on 1,261 German participants, each presented with vignettes about a fictional online news outlet. Participants were explicitly informed whether the content was produced by generative AI or by trained journalists. The authors varied two factors: the production process (AI versus journalists) and the subject matter (political versus entertainment). They then examined how participants responded: whether they trusted the content, whether they accepted the advertisements alongside it and whether they were willing to pay for the content. In short, the study explored not only how audiences perceived the journalism, but also how they sustain it through attention, advertising and financial support — demand in its fullest sense.

Encouragingly for journalists, the findings show the public places less trust in outlets using AI-generated content than in those publishing work by trained journalists. Acceptance of advertising — still a crucial revenue source for the great majority of news organizations — was also lower for AI-driven outlets. Regarding the willingness to pay for news, the authors did not find a difference between the AI and the journalist outlet. And, at least for these German participants, AI-generated political news was met with more skepticism than AI-generated entertainment content.

These results suggest that attempts to boost advertising revenue by producing more AI-generated content could backfire, since audiences are less willing to accept ads alongside AI-driven news. Taken together, the findings show that, when it comes to trust and advertising acceptance, these participants clearly value outlets that rely on human journalists. In other words, while AI may serve as a tool for journalism, it is not a replacement for journalism. Audience members seem to continue to favor the human touch, a key finding from this research that journalists would do well to emphasize. As the authors note, “In a world in which AI becomes ubiquitous in many sectors of society, traditional media organizations may benefit from emphasizing the ongoing role of journalists in communication to both current and potential consumers.”

Predatory practices — those in academic journals that prioritize profit over rigorous editorial and publication standards — are complex. These practices resist any simple predatory-versus-non-predatory dichotomy. Yet they remain well-known to scholars — charted, scrutinized, and generally avoided.

But journalists, who rely on academic journals for their reporting, are far less aware — or concerned — of this complex landscape, according to new research. The journalists exhibit a striking third-person effect, viewing predatory journals as a problem for other journalists but not something they themselves would fall victim to. At the same time, they have limited understanding of how their own instincts and decision-making processes might inadvertently narrow the range of research they share with the public.

Despite the important role health, science and environmental journalists play in translating academic research for the public, they are rarely recognized as “key stakeholders” in broader conversations about predatory journals. This study addresses that gap through 23 in-depth interviews with journalists from Europe and North America. It examines how they perceive predatory journals, how they determine whether a journal is trustworthy, and the implications of these perceptions and practices for the research that ultimately appears in the news.

Ultimately, the journalists interviewed relied on their own professional judgment — applying the same critical scrutiny they use for any other source — to evaluate an academic journal’s “trustworthiness.” Their decision-making processes depended heavily on assessments of a journal’s reputation and familiarity, as well as close attention to the professionalism of the articles it published, including clarity, grammar, structure, and overall presentation. This indicates that decisions about which research to feature are shaped less by formal measures of journal quality and more by the individual journalist’s experience, instincts, and interpretation — often exercised under tight deadlines, increasing daily demands, and rising professional precarity.

While turning to well-known, mainstream journals may feel safe, it comes with unintended consequences. Prioritizing reputation and prestige narrows the scholarship presented to the public, leaving work from lesser-known outlets — including regional, open-access, and Global South journals — largely invisible. The journalists interviewed often seemed unaware — and at times indifferent — to the ways their sourcing choices could perpetuate inequalities in which research reaches the public.

As predatory practices continue to evolve — becoming increasingly sophisticated, and in some cases even using AI-generated outreach to target journalists — it is more important than ever for reporters to exercise caution. Professional judgment and experience remain essential, but these instincts can be reinforced through structured strategies, such as interviewing journal authors or cross-checking sources, to verify the credibility of the research they feature. By pairing careful evaluation with intentional sourcing, journalists can safeguard the integrity of their reporting and protect public trust in science, health and environmental journalism.

The comedic news genre has never commanded more attention — or stirred more debate.

In the summer of 2025, CBS announced the cancellation of The Late Show with Stephen Colbert, just days after Paramount, its parent company, reached a multi-million-dollar settlement with Donald Trump. (Colbert won his first Emmy only two months after the announcement.) Last month, ABC suspended Jimmy Kimmel following a monologue about Charlie Kirk, before reinstating him amid public outcry. Although these flashpoints provoked serious concerns about freedom of expression, they simultaneously highlighted a vital reality. As trust in traditional journalism erodes, these comedic programs increasingly play an important role in informing citizenry.

It is against this backdrop that a team of researchers published an experimental study to explore how humor shapes viewers’ learning. Focusing on Last Week Tonight with John Oliver, the study examined how the placement of jokes affected the encoding of factual information. Fifty-seven participants watched the same 21 selected snippets, totaling just under 30 minutes. Using a recognition task alongside psychophysiological measures, the researchers tested how the proximity of facts to jokes — defined as within three sections before or after — impacted participants.

Ultimately, the study found that proximity matters. Participants showed heightened attention — measured through slowed heart rate — when facts immediately followed a joke. Further, humor appeared to enhance the “encoding” of facts presented near jokes. Encoding is the process of using our perceptions to create representations in our brains, thus helping us to pay attention, perceive, understand, and remember. The authors found that humor both before and after facts improved encoding, particularly when the facts and neighboring jokes were closely related.

These findings carry clear implications for the comedic news genre. As such programs further evolve to serve not only as sources of entertainment but also as vehicles for news and information, the way facts are presented becomes increasingly important. For producers, placing factual content alongside related jokes — especially in close proximity — may boost both audience attention and retention. This research suggests that humor, when carefully paired with information, does more than just make people laugh — it helps them remember information.

Content creators — from journalists to marketing agents to political strategists — labor over every word, weigh every phrase, and craft every sentence to deliver the perfect message. Yet all that effort perhaps can be amplified in an instant — with a smile.

In this experimental study, 120 university students viewed a series of stimuli designed like television documentaries, to test how a presenter’s smile might shape viewer engagement, emotional mimicry and evaluations of credibility and effectiveness. The stimuli featured content popular among students and focused on non-emotional information. Professionally trained actors delivered the same remarks twice, once with a neutral expression and once with a smile. The researchers tracked audience reaction through multiple methods: physiological indicators such as heart rate and skin conductance, automated facial coding to capture emotional mimicry, and participants’ self-reported ratings of the speaker’s credibility and effectiveness.

The study found a clear pattern: When presenters smiled, audiences responded more positively across all measures. Participants showed signs of greater attention and arousal through physiological measures, and their faces were more likely to mirror the presenter’s facial expression, as if smiling was contagious. Perhaps most notably, a simple smile boosted how credible and effective the presenter was perceived. Conversely, a neutral-faced presentation offered no spark of connection and, at times, produced the opposite effect, evoking mild feelings of anger or contempt.

These findings offer a compelling reminder for content creators — words alone are not enough. Audiences respond not just to what is said but how it is conveyed. Even a skillfully constructed script risks diminished impact if paired with a detached demeanor. A smile, however, signals warmth and credibility, deepening engagement in ways a script alone cannot. Still, not every context calls for a smile. Though not examined in this study, it is easy to see how audiences expect authenticity — and how, in moments of gravity such as dire news reporting, a smile may feel misplaced. Yet in today’s hyper-competitive attention economy, a fleeting spark of connection may be what determines whether content is ignored, remembered and shared.