Dodds, T., Mine, N., Helberger, N., Guzman, A. L., & Diakopoulos, N. (2026). AI hype in journalism: Visibility, power, and the politics of media narratives. Digital Journalism, 14, 207–219. https://doi.org/10.1080/21670811.2026.2630187
Summary
This editorial introduces a special issue of Digital Journalism that reconceptualizes AI hype not as a passing discursive bubble but as a systemic, structuring force within journalism. The authors — Dodds, Mine, Helberger, Guzman, and Diakopoulos — define AI hype as a set of exaggerated promises and anticipatory narratives that mobilize resources, redistribute legitimacy, and configure sociotechnical futures before they materialize. Drawing on infrastructure studies and community-of-practice theory, they argue that journalists occupy a contradictory dual role as both hype makers and hype watchers, and that debunking alone is insufficient. Instead they call for “breakdown-and-repair” approaches that render hype’s mechanisms visible and open space for alternative futures.
Key Contributions
- Offers a working definition of AI hype as a structuring sociotechnical force rather than mere discourse or transient expectation.
- Reframes the scholarly question from whether hype is justified to how it functions, mobilizes resources, and reconfigures journalistic labor.
- Synthesizes special-issue contributions into an integrated framework spanning communities of practice, recurring narratives, reach and scope, and breakdown/repair.
- Bridges journalism studies with STS infrastructure scholarship by applying “infrastructural inversion” to AI hype.
- Provides concrete recommendations: diversify expert sources, practice active reflexivity, and produce critical scholarship on labor, environmental, and societal impacts.
Methods
An editorial synthesis and theoretical framing rather than an empirical study. The authors draw on Star’s ethnography of infrastructure and Lave & Wenger’s community-of-practice theory, and conduct an integrative review of the special issue’s empirical studies — including ethnography of the Associated Press AI initiative, interviews with Chinese journalists, comparative content analyses across US/Dutch/Brazilian/German/Chilean/African media, and analyses of media-union responses to generative AI. Comparative historical analogies with prior hype cycles (Hyperloop, metaverse) contextualize the argument.
Findings
- Newsroom actors strategically leverage AI hype to secure funding and legitimacy — e.g., reframing older automation projects as “AI” projects.
- Sourcing has shifted from government officials in the “automation” era to tech executives once “AI” became dominant, sustaining hype and rendering alternative actors invisible.
- Media unions have negotiated GenAI protections, but vague language reproduces the invisibility of marginalized media workers.
- Journalistic discourse frames AI through “open-ended technological inevitability,” foreclosing questions about whether such futures should arrive at all.
- National framings diverge: Chinese journalists position as “loyal facilitators” aligned with state strategy, while German coverage emphasizes “European values” against US/China competitors.
- Hype obscures hidden labor (data annotators, moderators), environmental costs, and material infrastructures, and masks contradictions like labeling systems “fully automated” while requiring constant human supervision.
Connections
This editorial’s concern with how professional and platform actors construct legitimizing narratives around new technologies resonates with platform-critique work such as Helmond2026-ll and Baym2026-tr on the political economy and evolution of digital infrastructures. Its treatment of hidden labor, material infrastructure, and infrastructural inversion connects to Vertesi2026-lv, while its attention to how media narratives redistribute visibility and power sits alongside broader concerns about information disorder and public understanding of emerging technologies. Beyond these, the special-issue framing is largely distinct from the misinformation-focused papers under these topics.
Podcast
A research-radio episode discusses this paper: 🎧 MP3 · Spotify · Apple Podcasts