Body
ProPublica is at the center of a growing labor and governance debate over artificial intelligence in newsrooms, as its journalists and their union seek contract language on how AI should be used and what humans must control. Negotiations and “independent picket lines” earlier this month reflected the issues as the parties move toward an agreement that could include a strike, a prospect the union said would be driven by AI as the “chief sticking point.”
The dispute comes amid wider industry examples that journalists and media observers say have elevated urgency. The Associated Press story pointed to cases including corrections for errors in AI-generated summaries, removed articles tied to a fake author, and controversies involving fabricated text and unclear reader disclosure about when AI tools were used.
Union leaders said the aim is to secure commitments that mirror what the NewsGuild-USA has sought elsewhere in the industry about disclosure and the role of humans when AI is involved. Jen Sheehan, spokeswoman for the New York Guild, said members pledged “overwhelmingly that they would be willing to strike without a satisfactory agreement” during the contract talks.
Sheehan and others in the story framed the moment as part of a broader negotiation over who bears responsibility when AI drafting or assistance goes wrong. Alex Mahadevan, described as an expert at the Poynter Institute journalism think tank, said, “It feels to me pretty monumental when we think about the trajectory of AI and journalism,” underscoring how quickly newsroom practices are changing.
ProPublica, according to the story, has rejected the union’s requests. Tyson Evans, a spokesman for ProPublica, said the organization is “exploring how technology can create more space for investigative reporting,” and he discussed plans for expanded severance packages “in the ‘unlikely event’ of AI-related layoffs.” Evans said the contract should not freeze editorial decisions for too long, saying, “We’re approaching AI with both curiosity and skepticism,” and “It would be a mistake to freeze editorial decisions in a contract that will last years.”
The story also describes how widespread AI language is becoming in union contracts. Jon Schleuss, president of NewsGuild-USA, said “57 of 283 contracts” negotiated by the union “contain language related to artificial intelligence,” and he said the first such deals happened in 2023, with the Associated Press among early pioneers. Schleuss said the union wants similar provisions added to more contracts, arguing that accountability should stay with journalists rather than being outsourced to automation.
Beyond collective bargaining, the Associated Press story described disagreements about how much newsrooms should disclose AI use and how those disclosures affect audience trust. Benjamin Toff, director of the Minnesota Journalism Center at the University of Minnesota, said research shows that “a vast majority of American consumers” believe it is “very important that newsrooms tell the public when AI is used” in stories or images, and he discussed a finding that “a significant minority — 30% in a study” doesn’t want AI used at all.
Poynter’s Mahadevan raised a further complication, saying that disclosure rules can be hard to apply consistently because there are “just so many, many uses of AI in journalism, from the very beginning of the reporting process to when you hit publish.” He also said “just broadly declaring” that AI use in newsgathering requires disclosure can, in some cases, be “a disservice to the reader.” The story also noted that two New York state lawmakers introduced legislation requiring clear disclaimers when AI is used in published content, with no immediate indication of whether it would pass.
The AP story added examples of contrasting newsroom practices and attitudes toward AI. Chris Quinn, an editor described for The Plain Dealer in Cleveland, said his newspaper has sent journalists to gather quotes and information and then feed it to a computer to write, with human editing still part of the workflow. The Associated Press story said Quinn defended the approach as a way to make limited resources stretch, while also describing public arguments tied to how AI affects journalism careers and training.
In discussing what disclosures should mean in practice, the story quoted Mahadevan as saying the newsroom of the future “is going to look completely different than it does today,” and that “people will lose jobs” while new roles may appear. He said those changes make it important to hold conversations now because “audiences do not want a newsroom completely taken over by AI.”