The tensions inside the AP — and Rinehart’s articulation of a case many managers believe but are reluctant to make — reveal a broader conflict playing out across the media over how AI should be applied within journalism, a costly craft filled with strong-willed individuals.
Many (though notably not all) media companies have cut deals with AI giants to license their content, arguing that they would rather be compensated for their work than let it be scraped for free, and want to provide the models with quality information rather than digital garbage.
Within newsrooms themselves, media companies are rushing to adopt tools many of their employees are wary of using.
Most rank-and-file journalists, like many other white-collar workers, view AI tools with deep suspicion and see their adoption as potential (or inevitable) threat to their livelihoods.
Among newsroom and media leaders, the feeling is friendlier.
Axel Springer CEO Mathias Döpfner told me onstage last week at Semafor’s Trust in Media event that he was confident in the media business “because of the opportunities that AI can provide.”
In private, he’s been more direct. In a series of recent meetings with staff from Business Insider and Politico shared with Semafor, Döpfner has repeatedly emphasized that news organizations that do not embrace artificial intelligence will almost certainly be left behind and fail.
For the moment, news media seems more insulated than other professions from some of the threats of AI. Journalists spend years developing contacts and sources for information that can’t be gleaned or acquired in any other way than trust between two individuals.
Rinehart’s vision for AI tools is commonly held among some of her peer set, but seems to focus on the thorniest and, for the moment, least useful AI applications.
Many media companies have already developed fairly uncontroversial AI applications that have been broadly embraced precisely because they help journalists do what they can’t do at the moment.
In 2024, Semafor experimented with a partnership with Microsoft on a news aggregator called Signals, which helped find stories written in non-English languages not easily picked up by English searches. Earlier this year, Nieman Lab reported that The New York Times had built an AI podcast summarizer for its staff to better monitor the massive volume of content created every day across that fragmented ecosystem. Organizations like the Times and The New Yorker offer audio versions of originally reported stories narrated by AI. Transcription services have already made a formerly tedious process instantaneous, allowing reporters to process more information and get stories out quicker.
These tools are additive, and while they have the same potential for hiccups that other LLMs do, they are helpful at allowing journalists to access more and different information than was available before. Media companies are better off thinking of ways to capitalize on what AI already does well, particularly mass summarization, research, and visualization, rather than “helping” a few proud writers save a few minutes on their copy.