The Dynamics of the AI Rollout in Newsrooms

The next speaker in this AoIR 2024 conference session is Nadja Schaetz, whose interest is in AI hype in news coverage. Journalism has often uncritically covered the rise of generative AI, and swallowed the claims of AI companies about the capacities of their tools; this project collaborated with the Associated Press Local AI Initiative and conducted participant observation in local newsrooms to understand journalistic reactions to this initiative. Through the project AP worked with five newsrooms to provide AI-supported technologies.

What the study observed was not AI hype as such, however: not simply a gap between expectations and reality of artificial intelligence, but rather lack in the knowability of the truthfulness of AI claims. Participants also had many varying expectations, and it became more useful to unpack the sociology of expectations here. Early on in the Local AI initiative, building expectations about AI was crucial: this attracted newsroom participants and support from the AP leadership.

Newsroom professionals expected that AI use might give journalists more time to do other work, and such expectations generated interest and mobilised AP resources – AI Webinars were very well attended, for instance, and AP’s AI programme managers were also promoted in their jobs. However, such expectations must eventually also be met, and here news workers soon found that AI technologies still require considerable human involvement. AI does not necessarily simplify processes, at least in the experimental phase of the technology rollout – and success continually remains just out of reach.