Conference Agenda

Abstract

Building on crisis framings from journalism and information policy studies (Siles & Boczkowski, 2012), this article offers a new historical model linking technological disruption of news production, packaging, and distribution to a set of emerging policy responses. Whereas the web began to disrupt the packaging of news in the 2000s, social media platforms began to destabilize news distribution in the 2010s. In the mid-2020s, generative AI (GenAI) challenges content production itself, potentially undermining the sustainability of newsrooms and their capacity to support public deliberation in democratic systems.

We conceptualize crisis as a deep socio-technical destabilization. The economic crisis within the press sector is evident from various perspectives: in the United States, a pronounced decline in newspaper circulation has been accompanied by a significant reduction in newsroom employment—a decline of over 60 percent in revenue and circulation, coupled with a reduction of over 50 percent in newsroom employment (Pew Research Center, 2023).

The first crisis was fueled by the mass adoption of the web, whereby paper and the printing press were replaced by digital websites (Doyle, 2013). The second crisis was fueled by the mass adoption of social media networks, which had a direct impact on the distribution element and, with it, on the ability of newspapers to reach their readers. By the mid-2020s, the transition to the third crisis is already underway. The key shift is GenAI, which challenges the press's control over content production itself. In addition to an ever-growing number of news producers (a quantitative shift), content is increasingly being (re)produced by machines (a qualitative shift). Content production is no longer just labor, but capital. As Pickard (2023) argues, structural crises of this kind require systemic responses if journalism is to continue serving the public interest.

This article focuses on three critical challenges GenAI poses to journalism and the press. First, by privileging actors with access to computing power and data, it amplifies inequalities in content production. The high costs associated with data processing and energy consumption create steep market entry barriers, further consolidating control over information production and circulation in fewer hands (Crawford, 2021). Second, by enabling synthetic interactions at scale, it can erode interpersonal trust and dialogic practices necessary for the formation of publics (Ananny, 2018). GenAI amplifies existing concerns about inauthentic behavior because it can mimic specific personas and produce targeted content at lower cost and higher speed (Ferrara, 2023). Third, by introducing a competing mode of production based on statistical plausibility, it challenges source-based verification. Whereas journalism attributes factual claims to sources such as experts, authorities, and eyewitnesses, GenAI creates content based on statistical patterns emerging from a corpus, making provenance and verification harder.

In this article, we identify and analyze emerging policy responses organized around fair access, human agency, and transparency, and discuss their implications for newsrooms. On fair access, the EU funded the Trusted European Media Data Space (TEMS) to ensure that media organizations can safely share data into larger lakes and leverage such resources to develop bespoke AI systems (Prug & Bilić, 2024). Regarding human agency, a review of 37 AI guidelines from media organizations across 17 countries shows the call for human-in-the-loop practices is amongst the most consistent across different contexts (de-Lima-Santos et al., 2025). On transparency, there is a growing consensus that disclosure of training sets is necessary, be it through model cards or nutrition label styled approaches, with emerging practices including greater traceability of sources and edits.