Skip to content

Ohio journalists withhold bylines over DeeperDive AI summaries

Reporters at two USA TODAY Co. newspapers say the chatbot lacks human oversight and threatens editorial integrity.

Unionized journalists at two Ohio newspapers will withhold their bylines for 14 days beginning Oct. 28, protesting the addition of USA TODAY Co.’s DeeperDive AI tool to their websites. The action at the Akron Beacon Journal and Kent Ravenna Record-Courier spans the November general election. The journalists want their names removed from stories featuring DeeperDive, which generates AI summaries of the newspapers’ reporting. They are demanding that USA TODAY Co., which owns both titles, pause the tool and comply with its stated AI policies. The NewsGuild said the company requires a human in the loop before computer-generated content is published online or in print, while DeeperDive is being placed on stories without human oversight. The workers say the chatbot has hallucinated facts and breached editorial standards since its rollout across the company’s U.S. publications. In one October example cited by the union, the system replaced a murder suspect’s name with that of the reporter covering the case. “We don’t want our names on content that doesn’t meet common standards of editorial integrity,” Tawney Beans, chair of the Akron NewsGuild unit, said. DeeperDive was created by Taboola and was fully implemented for USA TODAY’s audience of more than 195 million monthly unique visitors in September 2025, after a beta launched in June. Gannett has said the engine draws on journalism from USA TODAY and the USA TODAY Network to answer reader questions and surface related reporting. It described the beta as being quality-assessed by humans and presented the product as a response to generative-AI search services that use publishers’ content without licensing or compensation. The Ohio dispute follows a broader pattern of newsroom resistance to AI summaries and search features. As covered in August, the New York Times Guild objected to the Times’ test of AI-generated search summaries, arguing that it crossed the paper’s AI guardrails.