The AI turn in journalism: Disruption,adaptation,and democratic futures

Abstract

In this essay, we argue that, unlike previous changes in digital media technologies over the past few decades, this AI “turn” in journalism forces us to rethink journalism’s identity and its relationship with audiences. While AI complicates and challenges some existing professional, social, political, and economic structures, it also offers new ways to realize desired journalistic objectives that were previously considered to be impractical, if not impossible. Drawing on four orienting ideas—adoption and hype, power and dependency, audiences and democratic implications, and education and empowerment—we unpack the implications of this AI turn in journalism and the consequences for the future of the journalistic field.

Keywords

Artificial intelligence automated journalism AI turn journalism robot journalism

For three decades, the story of journalism and digital technologies has been one of hamster-wheel acceleration, marked by increasing expectations to do more with less—to produce more news, participate on more platforms, and reach more diverse audiences, all with shrinking revenues and resources amid the general contraction of the legacy business model for news (Posetti, 2018). As Usher puts it, the “hamster wheel is a metaphor for news production in the digital age, where speed is more important than fact-checking, and quantity is more important than quality” (2014, p. 14).

But now, in the mid-2020s, the rapid rise of disruptive AI, particularly in the form of generative AI, offers a constitutive moment—perhaps even a “breaking point”—for journalism and technology: a junction where journalists and other newsworkers are forced to confront existential questions about their roles, routines, and relationships (Thomas and Thomson, 2023; Lewis et al., 2025a). The impacts will also be felt by news users who must navigate an onslaught of information (Peña-Fernández et al., 2023), raising fresh questions about how people engage with and make sense of news and information more broadly. In a world where nearly anyone can mass-produce content that, at minimum, appears to be of decent quality and seems plausible, what will it mean for people to determine what’s authentic, relevant, and valuable?

For some journalists, developments in artificial intelligence will only further accelerate the hamster wheel, leading to even greater demands to do more with less. Likewise, for some news consumers, the proliferation of AI will only further complicate their ability to puzzle through what’s happening in the world around them and exacerbate problems of “burnout” with news, with even more people experiencing frustration as they consume journalism or avoid it altogether (Borchardt, 2022). However, this critical moment also offers a generative opportunity of its own: an off-ramp to escape the hamster wheel, break away from tired routine, re-evaluate what it means to do good work and stay informed, and, overall, to more thoughtfully apply technology where it can create value in journalistic work and resist or refuse it where it cannot. For example, AI is already being used to make more content broadly accessible, whether through the creation of realistic audio renderings of text stories or the rapid creation of more accurate translations of articles (Arguedas and Simon, 2023).

To be sure, not all news organizations use AI. However, a broad and growing cross-section of them do―from local newspapers to public service broadcasters to large digital-native outlets. For example, 73.8% of respondents to a survey of journalists and media workers across six continents indicated that they or their organizations had already used generative AI in some capacity (Diakopoulos et al., 2024). The most common application was text-based content production, though it is also being widely used for information gathering and sensemaking, business applications, and coding. These findings align with a previous survey wherein news organizations from around the world reported being motivated by AI’s promise to improve journalists’ efficiency, deliver more relevant content to users, and enhance business efficiency (Beckett, 2019). Only about half of the organizations declared themselves AI-ready at that time, while the rest were either in the early stages of adoption or planning integration. While those strategies and integrations have undoubtedly matured between those surveys, they remain far from being coherently solidified within a normative professional framework.

Therefore, we argue in this essay that this “AI turn” in journalism forces us to rethink journalism’s identity and its relationship with audiences and provides an opportune moment to act on it. While AI complicates and challenges some existing professional, social, political, and economic structures, it also offers new ways to realize desired journalistic objectives that were previously considered to be impractical, if not impossible. Drawing on four orienting ideas—adoption and hype, power and dependency, audiences and democratic implications, and education and empowerment—we unpack the implications of this AI turn in journalism and its consequences for the future of the field.

The AI turn in journalism

Artificial intelligence (AI) is not just the future of journalism; it is part of its quite complicated present. Journalists around the world already regularly use tools and systems that can be fairly characterized as AI in their day-to-day work (Dodds et al., 2025), and managers of journalistic outlets are making decisions about the future of their organizations with AI in mind (Simon, 2024). While changes intricately tied to technological development aren’t novel in the news industry (Hermida and Young, 2021), the recent explosion of attention to AI is unusually wide-ranging (across industries), pervasive (across aspects of newswork), and fundamental (challenging traditional divisions between humans and computers). Consequently, AI appears to be diffusing within journalistic spaces in a way that has notable distinctions from past technologies—in part because journalists for years have expressed an unusual level of excitement for using AI to accelerate and augment their work, even as they expressed some unease about the existential threat it could pose for the nature of producing media (Lindén, 2017). Indeed, developments in AI—and generative AI, in particular, because of its capacity to produce humanlike content at scale in ways that threaten some of journalism’s core creative capabilities (Guzman and Lewis, 2024; Lewis et al., 2025b)—present a pivotal moment in the future of news, and one that may shape what counts as journalism, how news is made, and how people access and make sense of it.

To illustrate what makes increasingly capable AI, and generative AI in particular, an especially noteworthy development, consider how it can embed itself between producers (e.g., journalists) and users (e.g., audiences). AI systems can act as interpreters of information by summarizing and synthesizing journalistic content, with the journalist’s own words consequently never reaching the audience (Ufarte-Ruiz et al., 2023). AI can act as a conversational agent (Resendez et al., 2023), like chatbots that stand in for human journalists. AI can at least partially offload the responsibility of interacting with audiences, such as by suggesting personalized but canned responses to reader messages. In all these instances, AI introduces an entirely novel intermediary that further distances prototypical newsworkers from news users in important ways, even as it potentially brings news users closer to news products through increased relevance or usefulness.

The turn to AI-based systems in news organizations signals more than a trend. In academic and professional fields, a “turn” signifies a shift in the focus, paradigms, or methodologies, often driven by a broad societal change or the proliferation of a technological leap. In journalism studies, previous turns have marked transformational moments for the profession and its practices. For example, scholars pointed to an emotional turn in journalism that challenged the long-standing belief that journalism is primarily a rational enterprise, underscoring that emotions play central roles in shaping how modern journalism is produced, consumed, and understood (Wahl-Jorgensen, 2020). They have also pointed to a data turn spurred by the rise of computationally infused forms of journalism and folk theories about the rationality of “unbiased” data, giving new weight to data analysis, visualizations, and evidence-based narratives (Gray et al., 2012; Zamith, 2019). Lastly, scholars have pointed to an audience turn that placed publics and ‘users’ at the center of the journalistic process, supported by evidence of professionals increasingly seeking to understand audience behaviors, preferences, and interactions to engage more effectively with news users in a digital-first landscape (Costera Meijer, 2020).

While it can be easy to dismiss these turns as fleeting disruptions, we argue that they are more fundamental: they eventually become integrated into the fabric of journalistic practices, so much so that they become seen as natural components of the profession rather than revolutionary shifts or hype trends. At their inception, each turn highlighted tensions and prompted debates about how journalism should respond to new realities. Over time, however, these debates give way to consensus or adaptation, and the turns become the new normal.

The AI turn in journalism represents a similar moment of transformation and raises a corresponding question: Is this particular shift transitory or lasting? Like the aforementioned turns, the AI turn seems poised to embed itself within journalism’s norms and routines. As artificial intelligence tools and processes become more ubiquitous, they may lose their novelty and instead be seen as just another way of doing journalism. Yet, while the integration of AI in practice may eventually feel routine, its broader impacts—on labor, ethics, and democratic engagement—are likely to leave a lasting imprint on the field.

To that end, it is important to recognize that the AI turn in journalism is in part a continuation of longer-term trends in the field. The turn builds on the technological momentum created by previous developments, such as audience analytics, news recommendation systems, and the digitalization of journalism more broadly. None of the previous turns in journalism studies are independent from history or even from each other. Moreover, the AI turn shows evidence of becoming an accelerator for some of those trends. However, the possibility of using AI to reconfigure fundamental aspects of journalism―from the who to the what to the why to the how―coupled with the myriad concerns about the status quo of journalism means that phenomenon could well become a tipping point for reimagining journalism in a profound way.

Recognizing the AI turn as such helps us understand its significance. It is not simply a new technology to be adopted or a passing experimentation phase. Instead, it is a critical juncture that demands reflection and adaptation, much like the emotional, data, and audience turns before it. Although it is impossible to predict the future, and the tangible impact of these systems remains uncertain, we present evidence in this essay that shows that significant changes are already underway.

Hype and adoption

AI-infused journalism will be better and worse simultaneously, and in ways that only vaguely come into view as we see generative AI’s early sprouts. However, much of the emphasis to date has been, understandably, on AI’s ability to replace journalists or automate existing processes (Yerushalmy, 2023). Although this line of thinking has value, linking AI to its ability to replicate existing ways of thinking and doing misses the technology’s broader transformative potential. Current applications of AI are already much better than humans at performing some high-order tasks, even as they are considerably worse at other low-order tasks (Yeung and Dodds, 2024). This is because the state-of-the-art technology driving AI development fundamentally differs from how humans are wired (see Ramponi, 2023). Thus, the current focus on correspondence and replacement must be partnered with a meaningful consideration of complementarity and synergy.

The AI turn in journalism opens up a broader view that recognizes more disruptive possibilities resulting from complementarity and synergy, including AI’s potential to reconfigure (journalistic) roles, (professional) routines, (power) relations, and (news) experiences—with broader consequences for the rational-deliberative and consensus-building ideals typically associated with the notion of public spheres. However, disruptive possibilities also raise the prospect of undeserved hype, and there is little question that AI has become central to an ongoing ‘gold rush’ permeating a range of industrial sectors (Vrabič Dežman, 2024). Hype is not specific to AI, and it is not uncommon for inflated expectations to emerge as potentially disruptive technologies enter the mainstream (Lewis et al., 2025b). But while it is important to critically examine the actual limitations of the technology, it is also crucial to recognize that hype is socially meaningful in the AI turn and that imaginations about it, however flawed, play consequential roles in decision-making—from how news organizations allocate increasingly limited resources to the ways they engage with other social actors. Indeed, journalists serve as tastemakers for society, contributing to the imagined futures of technologies, and a large-scale analysis of tweets before and after the launch of ChatGPT in late 2022 suggests that journalists’ emotions were more positive than not, potentially influencing the narratives that developed about generative AI chatbots (Lewis et al., 2025b).

It is also important to recognize that AI is already meaningfully embedded in contemporary journalism. As Schjøtt Hansen et al. (2020) argue, AI systems are being used today across the entire media cycle, from making archival content more accessible to automating note-taking and reducing the cost of transcriptions to copy-editing and summarizing content to erecting increasingly dynamic paywalls. In other words, it is sometimes harder to find aspects of journalism that have not yet been touched by AI (Simon, 2022). Likewise, as more journalistic functions become tied to AI—from AI-powered automation of technical work in broadcast control rooms to widespread adoption of generative models to rapidly produce aesthetic elements—the field will become further structured by tools and infrastructures that were not designed to serve journalism.

While there is good reason to be concerned about the negative impacts these changes may have on journalism—or what it might simply perpetuate—it is important to be mindful that technologies are not deterministic. AI’s technological affordances can and have been used to improve, and occasionally break from, contemporary practice. AI has been shown to be effective in streamlining repetitive journalistic tasks like transcription, translation, and archival research (Fridman et al., 2023). AI has helped journalists expand their reach across languages and formats (Canavilhas, 2022). It has provided news organizations relatively inexpensive tools to help them detect biases or improve diversity in their work (Shin et al., 2022). AI’s possibilities are not just hype because these are already-realized benefits within the field. There are therefore highly plausible futures wherein AI is used to offload the so-called ‘grunt’ work of journalism that would allow journalists to re-focus on the sorts of activities that professionals and citizens alike say they value most: in-depth investigations, on-the-ground reporting, and other core journalistic activities that have increasingly become sidelined in the never-ending chase of the hamster wheel. While such futures require, to some extent, a parallel break from the business logic that has fueled the hamster wheel in recent decades, the AI turn at least provides both already existing technology and an opportune moment to make this break more economically viable for actors that wish to pursue it.

Power and dependency

The AI turn is also marked by its threat to the business models underpinning journalism around much of the world, and the relationships it has developed with other social actors—and tech companies in particular (Helberger, 2020). While AI offers cost-saving opportunities through automation, it also amplifies the power held by tech platforms that supply the technology and that serve as intermediaries between news producers and users (Simon, 2023). Consequently, more of the public opinion-shaping power that had been associated—fairly and not—with traditional media organizations continues to shift to the companies and platforms that provide the data, technology, and infrastructures that journalists and their audiences use in the modern media environment (Dodds et al., 2023). The AI turn in journalism thus continues the fundamental rearranging of what used to be considered the ‘core’ and ‘peripheries’ of journalism, as well as critical dependencies that exist among those actors.

Moreover, the AI turn appears poised to accelerate the centralization of power among fewer organizations. This includes larger journalism incumbents whose scale gives them negotiating power as well as prominent non-journalistic intermediaries whose positions within information networks force (growing) dependencies. We have already seen the impacts of this broader trend on local news outlets in the U.S., which not only bore the brunt of economic disruption but also had to increasingly tie themselves to platform companies like Facebook and Google (Usher, 2021). Those very companies are now major players in the development and application of AI in journalism (Simon, 2022), exacerbating the challenges and deepening the dependencies.

For example, Google’s adoption of AI-based summarization in its search engine results is likely to further upend the economic foundation for journalism by significantly reducing referral traffic (Hagey et al., 2023). With time, those summaries will likely satisfy enough users to preclude clicking through to original sources (if they are even linked to at all). Not only will this further jeopardize ad-based business models but it will also reduce the brand value of many organizations as the context is further collapsed. While some media companies have already responded by licensing their content to companies like OpenAI, the value of the traffic losses is likely to exceed the licensing revenue. And, those that have opted to firewall their content as best they can are opting to take a considerable risk of becoming less relevant.

Different countries and regions have responded differently to this rearrangement of power. For example, the U.S. has mostly turned to voluntary ethical codes crafted by private companies and controversial “safety” units within them, with regulatory efforts shunned as impediments to innovation. On the other hand, China has actively pursued AI as a strategic interest and encouraged technological development, even when those advances come at the expense of its private media industry (Kuai et al., 2022). In contrast, the European Union (EU) has adopted a more multifaceted regulatory approach to address platform power and dependencies, with key initiatives such as the Digital Services Act (DSA), the European Media Freedom Act (EMFA), and the Digital Markets Act (DMA). However, some scholars argue that these initiatives still have critical blind spots, particularly when addressing digital media concentration and platform dependence (Seipp et al., 2024).

Ultimately, news organizations’ business and operational strategies are likely to become unavoidably further intertwined with the interests of tech and platform companies (Simon, 2022), requiring them to develop clear plans and guidelines for dealing with both the appropriate uses of AI technologies as well as the problematic interrelationships that arise. Alternatively, the AI turn may provide new impetus for news organizations to engage more closely with open-source communities to reduce some of those dependencies, especially in light of the fact that some powerful AI systems—from large language models to audio transcription tools—have been released under open-source licenses. Some of these collaborations have already emerged as part of movements toward open data and open source intelligence investigations (Ganguly, 2022), with the AI turn offering a juncture for more fundamental integrations.

Audiences and democratic implications

The AI turn provides the opportunity to reevaluate what is most desirable about—or needed from—journalism (Costera Meijer, 2022), which in turn provides the foundation for thinking about how AI can enable work toward those objectives. For example, Lin and Lewis (2022) offered a normative baseline when they argued that journalistic uses of AI should focus on supporting democratic wellbeing by improving the accuracy, accessibility, diversity, relevance, and timeliness of news. Indeed, if deployed thoughtfully in journalism, AI could enhance the value proposition of news in several ways. These include making journalism more accessible and readable (e.g., offering easy-to-understand summaries or providing “catch me up” distillations of previous news reports); reducing the expense required to produce news (e.g., lowering transcription and translation costs) and enabling more affordable news access for consumers; and developing information personalization tools that accentuate knowledge about local politics or civic engagement opportunities. As noted above, even as AI distances news users from producers, it can potentially make news products more relevant and useful to those users.

At the same time, the AI turn poses serious threats to the informational wellbeing of audiences and the functioning of democratic societies. It is now easy and relatively inexpensive to produce content at scale that has the appearance of being high quality (even when it is nothing more than an articulate-sounding mish-mash of ideas) and that can be optimized to draw attention from search engines and aggregators or for shareability. Companies like Copysmith, Jasper, and Kafkai have capitalized on the development of generative AI tools to create wrapper services that make it easy for anyone to quickly and inexpensively produce AI-generated and SEO-friendly tailored “content”—anything from marketing copy to press releases to social media posts. While these services are sometimes used to create drafts that professionals then refine, they are also used within algorithmic workflows in ways that harm journalism and its publics.

For example, generative AI tools can be used to sow confusion by mass producing disinformation (Bontridder and Poullet, 2021). In many instances, disinformation isn’t intended to get people to believe a particular thing but rather to promote conflicting perspectives that unsettle them (Hameleers, 2023). Likewise, the same tools can be used to mass-produce “AI slop,” or junk content and unauthorized derivatives often created with the aim of generating advertising revenue by drawing traffic from indiscriminate humans and bots (Braun and Eklund, 2019). Regardless of the motivation, that junk content ends up crowding out ‘quality’ information, making it harder for citizens to find it. And, as media organizations deepen their involvement with companies that operate AI systems, audiences face a looming “authenticity crisis,” struggling to distinguish between human-created and AI-generated content. Although efforts are underway to develop systems for watermarking AI-generated material and tracking the provenance of news, it is too soon to tell whether these measures will prove effective (Piasecki et al., 2024). In other words, news organizations aren’t the only (or even primary) users of the technologies at the center of the AI turn. It’s also available to trolls, profiteers, and nation-states. This raises a very real prospect for a recurring question: who’s to say that a given piece of (news) content isn’t junk?

While concerns about filter bubbles and information silos predate the current AI boom (Bruns, 2019), the shifts driven by generative AI no doubt exacerbate them due to the combination of sophistication and scale (see Shoaib et al., 2023). For example, it is now much easier for malevolent actors to use generative AI to produce deepfakes and automate cheapfakes, and then deliberately lead individuals down rabbit holes through the inclusion of AI-optimized keywords. Likewise, AI’s ability to profile and more intelligently discriminate, sort, and pander helps entrench epistemic bubbles that concurrently limit the exposure to alternative views and seeds distrust of outside sources. This leads Coeckelbergh (2024) to contend that “AI as it is currently used and developed endangers democracy” in part by “undermining the knowledge and trust basis of democracy” (p. 7).

It is easy to overestimate these fears and highlight AI’s potential to promote mass social manipulation. Without diminishing those prospects, we instead contend that the AI-facilitated pollution of information streams and the destabilization of audience trust will further promote news “burnout” and related frustrations. More importantly, it will likely lead to subsequent detachment—not only from information but also from social and political processes—as citizens struggle to deal with the cognitive load of ascertaining what is true, relevant, and valuable. Such burnout and detachment from the media environment is an element that, as yet, has not been examined as part of the broader phenomenon of news avoidance evident among many consumers worldwide (Toff et al., 2023). The AI turn, therefore, requires us not only to recognize but also plan a response to the pollution of mainstream information streams and the growing presence of malicious synthetic agents, all the while adjusting to the even more customized and personalized mediations of social reality that will characterize the AI turn in journalism in the coming years. To that end, computer scientists—sometimes in collaboration with journalists—continue to develop tools that leverage AI to help newsworkers and users alike spot and track disinformation and become more aware of their epistemic bubbles (Singh et al., 2024). While these developments are currently reminiscent of war—new defensive tools are developed in response to new offensive threats—the AI turn presents an opportunity to rethink what we want citizens to get out of journalism and promote partnerships between industry, civic groups, and educational institutions to create agents and tools that advance those ends (Coeckelbergh, 2024).

Education and empowerment

Journalism faces profound knowledge gaps and epistemic challenges amid the AI turn. Many journalists and educators in journalism schools understand the theoretical frameworks that guide journalism but lack technical literacy about AI systems. Conversely, researchers who study AI, many of them in computer science and related fields, often have limited familiarity with the ethical and professional challenges that define journalism. These knowledge silos—which exist within both the newsroom and the academy—hinder the development of integrated approaches that respect journalistic values and stifle efforts to incorporate new technologies effectively (Dodds et al., 2025). Intentional efforts to address knowledge gaps and silos around AI are therefore necessary as the AI turn further shapes the profession and society.

For educators in secondary schools and universities, the AI turn has already posed a profound disruption. Teachers are scrambling to reorient their assignments to be “AI proof” amid the proliferation of generative AI tools that can produce convincing software-made replicas of human reasoning and argumentation. For journalism educators in particular, these developments can feel doubly fraught: Beyond the increased possibilities of student cheating, the hallucinations of generative AI seem especially threatening to the fidelity toward facts that is sacrosanct in journalism education. Additionally, aspiring journalists have long been taught about the primacy of “original content”—original ideas, first-hand observations, exclusive interviews, etc.—as the core of journalism’s self-understanding as a profession. How much AI involvement in crafting and editing content, therefore, should be allowed before that human originality is compromised? At the same time, the jobs for which student journalists are being trained increasingly call for skills in using AI—which suggests that the news industry, perhaps once again, is running ahead of the academy when it comes to technological adoption (Guzman and Lewis, 2024). In other words, failing to familiarize students with AI tools shortchanges them both as citizens and aspiring professionals.

A recent study by Wenger et al. (2024) of 14 accredited journalism schools in the U.S. highlights this issue in finding that one “significant challenge for program administrators is the lack of deep faculty expertise on the subject of AI,” which may explain why none of the programs had yet developed “a comprehensive plan around AI instruction for journalism students” (p. 13). Instead, the question of whether AI will fit into journalism education (if at all) appears to be up to the individual faculty member in most cases. Notably, while all of the university administrators participating in the study agreed that a strong ethical orientation for the use of AI is essential, “the path to achieving that goal remains undefined” (p. 13). While ad-hoc responses to rapidly unfolding disruptions are not surprising, this lack of vision, coordination, and decisiveness on the part of universities is still problematic, and underscores the need for interdisciplinary scholarship of teaching and learning (SoTL) around AI that blends pedagogy, computer science, critical AI studies, human-computer interaction, and journalism studies. At the very least, the AI turn presents an opportune moment to not only rethink what journalism education should entail but to act on it through curriculum redesigns that offer more robust protections against the ongoing misuses of AI, even as they also programmatically integrate such systems as capable assistants—teaching students, for example, how to use AI as tutor, research assistant, and idea generator (e.g., see Mollick, 2024).

Addressing the changes wrought by the AI turn cannot be limited to education for future practitioners. Such efforts must also target existing professionals, whether through formal training programs or informal but intentional professional socialization efforts. Notable efforts have been launched by media organizations like The Associated Press and educational institutions like Polis, with tech companies like Facebook, Google, and OpenAI launching their own alternatives. Notably, the most prominent efforts are led by organizations based in the Global North, which raises questions about how such training serves to empower those organizations and further embed their processes and ways of thinking into journalism and society broadly (Salgado Arzuaga, 2022). Even when otherwise celebrated institutions or non-profits lead the efforts, it is essential to recognize and reflect upon their potential to colonize local journalism cultures (Cheruiyot et al., 2019).

Finally, we must turn our attention to non-professionals and citizens. The AI turn will demand a broader societal emphasis on media literacy and involvement from news users. Journalism’s role in shaping collective perceptions and empowering audiences to make informed decisions in a democratic context is well understood, even if the actual impact of journalism in media culture has arguably waned even before the AI turn as news becomes increasingly crowded out by the proliferating options for entertainment (Boczkowski, 2021) and as power shifted toward digital intermediaries (Nielsen and Ganter, 2022). The AI turn further complicates those developments by wresting more of the control away from professional journalists and empowering news users to further remix their experiences based on parameters they (or the digital intermediaries) set. These new powers come with added responsibilities.

Thus, we need further evaluation of journalism’s less-well-understood role (yet one that is more pressing than ever) as a media-and-technology educator for society. Amid the AI turn, that may mean offering people tangible tips and techniques for navigating an information and social environment made more complicated by a growing array of human-sounding chatbots, AI-generated fakery on social media, and black-boxed AI systems that call for the kind of investigative scrutiny and explanatory storytelling that journalists are uniquely trained to provide. Thus, at a time when journalism’s influences seem to be on the decline, journalists can reclaim a degree of relevance (Carlson et al., 2022). They can do so not only by improving their work through the affordances of AI but also by prioritizing work that, at least in part, serves to guide citizens through the uncertain landscape of AI-driven communication that lies ahead.

Footnotes

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Tomás Dodds

Rodrigo Zamith

Seth C Lewis

Author biographies

Tomás Dodds is an Assistant Professor in Journalism and New Media at Leiden University and a Faculty Associate at the Berkman Klein Center for Internet & Society at Harvard University. He is also a researcher in the AI, Media & Democracy Lab in the Netherlands and the Artificial Intelligence and Society Hub [IA + SIC] in Chile.

Rodrigo Zamith is an Associate Professor and Chair of the Journalism Department at the University of Massachusetts Amherst. His research and teaching interests lie at the intersection of journalism and technology, with a focus on the reconfiguration of journalism in a changing media environment and the development of digital research methods for social scientists. Zamith’s work may be found at .

Seth Lewis is Professor, Director of Journalism, and the founding holder of the Shirley Papé Chair in Emerging Media in the School of Journalism and Communication at the University of Oregon—a position he has held since joining UO in 2016. He has held visiting or affiliated fellow positions with the Tow Center for Digital Journalism at Columbia University, the Information Society Project at Yale Law School, and Oxford’s Reuters Institute for the Study of Journalism, among others. From 2020 to 2022, he served as the elected Chair of the International Communication Association’s Journalism Studies Division, the world’s largest scholarly group dedicated to the study of journalism.

References

Arguedas

Simon

(2023) Automating Democracy: Generative AI, Journalism, and the Future of Democracy. University of Oxford. https://ora.ox.ac.uk/objects/uuid:0965ad50-b55b-4591-8c3b-7be0c587d5e7

Beckett

(2019) New Powers, New Responsibilities A Global Survey of Journalism and Artificial Intelligence. JournalismAI, Polis. https://drive.google.com/open?id=1utmAMCmd4rfJHrUfLLfSJ-clpFTjyef1

Boczkowski

(2021) Abundance: On the Experience of Living in a World of Information Plenty. Oxford: Oxford University Press.

Bontridder

Poullet

(2021) The role of artificial intelligence in disinformation. Data & Policy 3: e32.

Borchardt

(2022) Go, robots, go! The value and challenges of artificial intelligence for local journalism. Digital Journalism 10(10): 1919–1924.

Braun

Eklund

(2019) Fake news, real money: ad tech platforms, profit-driven hoaxes, and the business of journalism. Digital Journalism 7(1): 1–21.

Bruns

(2019) Filter bubble. Internet Policy 8(4): 1–14. Available at: https://policyreview.info/concepts/filter-bubble.

Canavilhas

(2022) Inteligencia artificial aplicada al periodismo: Estudio de caso del proyecto “A European Perspective” (UER). Revista Latina de Comunicación Social 80: 1–16.

Carlson

Robinson

Lewis

(2022) News after Trump: Journalism’s Crisis of Relevance in a Changed Media Culture. Oxford: Oxford University Press.

10.

Cheruiyot

Baack

Ferrer-Conill

(2019) Data journalism beyond legacy media: the case of African and European civic technology organizations. Digital Journalism 7(9): 1215–1229.

11.

Coeckelbergh

(2024) Why AI Undermines Democracy and what to Do about it. Cambridge: Polity.

12.

Costera Meijer

(2020) Understanding the audience turn in journalism: from quality discourse to innovation discourse as anchoring practices 1995–2020. Journalism Studies 21(16): 2326–2342.

13.

Costera Meijer

(2022) What is valuable journalism? Three key experiences and their challenges for journalism scholars and practitioners. Digital Journalism 10(2): 230–252.

14.

Diakopoulos

Cools

, et al. (2024) Generative AI in Journalism: The Evolution of Newswork and Ethics in a Generative Information Ecosystem. New York, NY: The Associated Press.

15.

Dodds

de Vreese

Helberger

, et al. (2023) Popularity-driven metrics: audience analytics and shifting opinion power to digital platforms. Journalism Studies 24(3): 403–421.

16.

Dodds

Vandendaele

Simon

, et al. (2025) Knowledge silos as a barrier to responsible AI practices in journalism? Exploratory evidence from four Dutch news organisations. Journalism Studies 26(6): 1–19.

17.

Fridman

Krøvel

Palumbo

(2023) How (not to) run an AI project in investigative journalism. Journalism Practice 1–18. DOI: 10.1080/17512786.2023.2253797.

18.

Ganguly

(2022) The Future of Investigative Journalism in the Age of Automation, Open-Source Lintelligence (OSINT) and Artificial Intelligence (AI). Doctoral, University of Westminster, London.

19.

Gray

Chambers

Bounegru

(2012) The Data Journalism Handbook: How Journalists Can Use Data to Improve the News. 1st edition. Sebastopol, CA: O’Reilly Media.

20.

Guzman

Lewis

(2024) What generative AI means for the media industries, and why it matters to study the collective consequences for advertising, journalism, and public relations. Emergency Medicine 2(3): 347–355.

21.

Hagey

Kruppa

Bruell

(2023) News Publishers See Google’s AI Search Tool as a Traffic-Destroying Nightmare. The Wall Street Journal. https://www.wsj.com/tech/ai/news-publishers-see-googles-ai-search-tool-as-a-traffic-destroying-nightmare-52154074

22.

Hameleers

(2023) Disinformation as a context-bound phenomenon: toward a conceptual clarification integrating actors, intentions and techniques of creation and dissemination. Communication Theory 33(1): 1–10.

23.

Helberger

(2020) The political power of platforms: how current attempts to regulate misinformation amplify opinion power. Digital Journalism 8(6): 842–854.

24.

Hermida

Young

(2021) Journalism Innovation in a Time of Survival. In: Luengo

Herrera-Damas

(eds). News Media Innovation Reconsidered: Ethics and Values in a Creative Reconstruction of Journalism. Hoboken, NJ: John Wiley & Sons, 40–52. Available at: https://onlinelibrary.wiley.com/doi/10.1002/9781119706519.ch3.

25.

Kuai

Ferrer-Conill

Karlsson

(2022) AI ≥ journalism: how the Chinese copyright Law protects tech giants’ AI innovations and disrupts the journalistic institution. Digital Journalism 10(10): 1893–1912.

26.

Lewis

Guzman

Schmidt

Bibo

(2025a) Generative AI and its disruptive challenge to journalism: An institutional analysis. Communication and Change 1. DOI: 10.1007/s44382-025-00008-x.

27.

Lewis

Markowitz

Bunquin

JBA

(2025b) Journalists, emotions, and the introduction of generative AI Chatbots: a large-scale analysis of tweets before and after the launch of ChatGPT. Social Media + Society 11(1): 20563051251325597.

28.

Lin

Lewis

(2022) The one thing journalistic AI just might do for democracy. Digital Journalism 10(10): 1627–1649.

29.

Lindén

C-G

(2017) Algorithms for journalism: the future of news work. The Journal of Media Innovations (Discontinued) 4(1): 60.

30.

Mollick

(2024) Co-Intelligence: Living and Working with AI. New York, NY: Portfolio.

31.

Nielsen

Ganter

(2022) The Power of Platforms: Shaping Media and Society. Oxford: Oxford University Press.

32.

Peña-Fernández

Meso-Ayerdi

Larrondo-Ureta

, et al. (2023) Without journalists, there is no journalism: the social dimension of generative artificial intelligence in the media. Profesional de la 32(2): 1–15.

33.

Piasecki

Morosoli

Helberger

, et al. (2024) AI-generated journalism: do the transparency provisions in the AI Act give news readers what they hope for? Internet Policy 13(4): 1–28. Available at: https://policyreview.info/articles/analysis/ai-generated-journalism-transparency-provisions.

34.

Posetti

(2018) Time to Step Away from the ‘Bright, Shiny Things’?: Towards a Sustainable Model of Journalism Innovation in an Era of Perpetual Change. (pp. 1–30). Oxford: Reuters Institute for the Study of Journalism. Available at: https://doi.org/10.60625/RISJ-KMPG-Q993.

35.

Ramponi

(2023) The Full Story of Large Language Models and RLHF [AssemblyAI]. News, Tutorials, AI Research. https://www.assemblyai.com/blog/the-full-story-of-large-language-models-and-rlhf/

36.

Resendez

Araujo

Helberger

, et al. (2023) Hey Google, what is in the news? The influence of conversational agents on issue salience. Digital Journalism 1–23.

37.

Salgado Arzuaga

(2022) From the classroom to the newsroom: a critical route to introduce AI in journalism education. Facts and Frictions. https://factsandfrictions.ca/portfolio-item/ffv2n1-leslie-salgado/

38.

Schjøtt Hansen

Helberger

Blanke

, et al. (2020) Initial White Paper on the Social, Economic, and Political Impact of Media AI Technologies (No. D2.2). AI4Media - A European Excellence Centre for Media, Society and Democracy. https://www.ai4media.eu/reports/initial-white-paper-on-the-social-economic-and-political-impact-of-media-ai-technologies-2/

39.

Seipp

Helberger

De Vreese

, et al. (2024) Between the cracks: blind spots in regulating media concentration and platform dependence in the EU. Internet Policy 13(4): 1–25. DOI: 10.14763/2024.4.1813.

40.

Shin

Hameleers

Park

, et al. (2022) Countering algorithmic bias and disinformation and effectively harnessing the power of AI in media. Journalism & Mass Communication Quarterly 99(4): 887–907.

41.

Shoaib

Wang

Ahvanooey

, et al. (2023) Deepfakes, misinformation, and disinformation in the era of Frontier AI, generative AI, and large AI models. (No. arXiv:2311.17394). arXiv. DOI: 10.48550/arXiv.2311.17394.

42.

Simon

(2022) Uneasy bedfellows: AI in the news, platform companies and the issue of journalistic autonomy. Digital Journalism 10(10): 1–23.

43.

Simon

(2023) Escape me if you can: how AI reshapes news organisations’ dependency on platform companies. Digital Journalism 12(2): 1–22.

44.

Simon

(2024) Artificial Intelligence in the News: How AI Retools, Rationalizes, and Reshapes Journalism and the Public Arena. Tow Center for Digital Journalism, Columbia University. https://ora.ox.ac.uk/objects/uuid:aeb25013-1d17-40b2-b471-5bdca309db87

45.

Singh

Ahmed

Alam

, et al. (2024) A comprehensive review on automatic detection of fake news on social media. Multimedia Tools and Applications 83(16): 47319–47352.

46.

Thomas

Thomson

(2023) What does a journalist look like? Visualizing journalistic roles through AI. Digital Journalism 1–23. DOI: 10.1080/21670811.2023.2229883.

47.

Toff

Palmer

Nielsen

(2023) Avoiding the News: Reluctant Audiences for Journalism. New York, NY: Columbia University Press, 288.

48.

Ufarte-Ruiz

M-J

Murcia-Verdú

F-J

Túñez-López

J-M

(2023) Use of artificial intelligence in synthetic media: first newsrooms without journalists. Profesional de la 32(2): 1–17. DOI: 10.3145/epi.2023.mar.03.

49.

Usher

(2014) Making News at The New York Times. University of Michigan Press. https://www.press.umich.edu/5134391/making_news_at_the_new_york_times

50.

Usher

(2021) News for the Rich, White, and Blue: How Place and Power Distort American Journalism. New York, NY: Columbia University Press.

51.

Vrabič Dežman

(2024) Promising the future, encoding the past: AI hype and public media imagery. AI and Ethics 4(3): 743–756.

52.

Wahl-Jorgensen

(2020) An emotional turn in journalism studies? Digital Journalism 8(2): 175–194.

53.

Wenger

Hossain

Senseman

(2024) AI and the impact on journalism education. Journalism & Mass Communication Educator 80(1): 97.

54.

Yerushalmy

(2023) German Publisher Axel Springer Says Journalists Could Be Replaced by AI. The Guardian. https://www.theguardian.com/technology/2023/mar/01/german-publisher-axel-springer-says-journalists-could-be-replaced-by-ai

55.

Yeung

Dodds

(2024) Automated journalism. (No. arXiv:2409.03462). arXiv. DOI: 10.48550/arXiv.2409.03462.

56.

Zamith

(2019) Transparency, interactivity, diversity, and information provenance in everyday data journalism. Digital Journalism 7(4): 470–489.