• by Ray Schultz , June 18, 2023

Gannett, the largest newspaper publisher in the U.S., will introduce generative AI later this year in a careful deployment that will include oversight by humans. 

The company will roll out a pilot program that will use AI to identify the important points of a story and create buffeted summaries at the top in USA Today in the fourth quarter, Reuters writes. 

The program will also equire human vetting. 

“The desire to go fast was a mistake for some of the other news services,” says Renn Turiano, senior vice president and head of product at Gannett, according to Reuters. “We’re not making that mistake.”

Fears exist that AI will help big companies like Gannett further cut journalistic jobs. Last week, hundreds of Gannett staffers nationwide walked conducted a one-day walkout to protest what the NewsGuild union calls the hollowing out of newsrooms.  

Gannett is relying on Cohere, training its large language model on 1,000 stories with reporter-written summaries, Reuters continues. Cohere competes with the Microsoft-backed OpenAI.

For its part, Reuters is using AI for voice-to-text transcription to produce scripts and subtitles for video, but it is not publishing AI-generated stories, videos or photographs, the story continues.