Making Sense of the Synthetic Press: How the Journalistic Field Defines, Anticipates, and Manages the Risks of Generative AI
Mike Ananny and Jake Karr

Summary of unpublished article.

If the ways that journalists work and the stories that they tell always depend upon an era’s technologies, then what kind of news emerges from a press that relies on Generative Artificial Intelligence (GenAI)? How is this press actually a “synthetic press” that uses machine learning to create its words, images, and videos—and to shape its workflows, economics, ethics, intellectual property, and sense of public purpose? As in earlier eras of change with intertwined technological, social, political, cultural, and economic forces, today’s press is asking foundational questions: what is GenAI to journalism, why does it matter, and how can its risks and opportunities be managed in ways that realize press autonomy and public-service journalism?

Motivated by debates in neo-institutional sociology (how do fields define and change themselves?), science and technology studies (how do communities and controversies stabilize new technologies?), and media industry research (what creates the “networked press” and “hybrid media systems”?), this project traces how journalism grappled with GenAI as its popularity exploded in 2022–23. By analyzing news discourse about GenAI, news organizations’ GenAI policies, and interviews with leading journalists and media lawyers, we see a nascent “synthetic press” reacting to fast-paced change, making sense of errors and controversies, reflecting on its dependencies on technologies, imagining and safeguarding itself against different futures, and revisiting its existential economics and public missions.

The synthetic press certainly resembles early eras of journalism—celebrating the profession’s core investment in discovering and narrating factual stories in the public interest—but it shows early signs of significant divergence. It seems unsure of what exactly defines its work, with a willingness to delegate backend tasks and seemingly ancillary actions to machine learning infrastructures. It struggles with whether its intellectual property and revenue models should be protected from technology companies’ scrapers and data models, while grappling with the possibility that journalism’s commitment to creating and publishing verifiable facts might actually improve the veracity of GenAI systems for everyone. And it shows early signs of institutional cracks and fragmentations as it variously embraces or rejects GenAI, with some organizations, beats, reporting styles, and labor politics being more or less compatible with synthetic media’s infrastructures and industries.

Envisioned as the first in a series of investigations, this project teases out early dynamics of a field in transition. By analyzing discourses, policies, and practices defining this journalistic moment, we see a “synthetic press” emerging that is likely to continue shifting for years to come.