US newspaper giant pauses AI-written articles after social media mockery
- Gannett has paused the use of AI to write sports articles after readers pushed back.
- AI-written articles quickly went viral over their bizarre language and lack of detail.
A US newspaper giant has paused the use of AI to write articles after they were savagely mocked on social media.
Gannett, which owns hundreds of local newspapers in the US, had been experimenting with using AI-powered sports writing tool LedeAI in its regional publications. However, readers quickly pointed out that many of these articles were lacking in detail, badly written, and read as if the writer had never watched the sport in question.
A spokesperson for Gannett confirmed to Insider that the company had paused the use of LedeAI tools to write articles, a move initially reported by Axios.
"In addition to adding hundreds of reporting jobs across the country, we have been experimenting with automation and AI to build tools for our journalists and add content for our readers," they said.
"We have paused the high school sports LedeAI experiment and will continue to evaluate vendors as we refine processes to ensure all the news and information we provide meets the highest journalistic standards," they added.
Some of its AI-generated articles were mocked on social media for their repetition and jarring cliches. One report from the Columbus Dispatch described a game between Westerville North and Westerville Central as a "close encounter of the athletic kind."
In another AI-produced article, the team names failed to generate entirely, with the opening line reading "The Worthington Christian [[WINNING_TEAM_MASCOT]] defeated the Westerville North [[LOSING_TEAM_MASCOT]]."
The stories were changed after they went viral on social media, with the Dispatch adding a note that they had "been updated to correct errors in coding, programming or style."
Insider identified AI articles in a range of Gannett's other local publications, including the Des Moines Register, the Arizona Republic, Florida Today, and the Milwaukee Journal Sentinel. All feature repetitive and robotic language.
Gannett is among a number of publications that have experimented with AI-generated journalism, with mixed results. The tech publication CNET scrapped its own AI-writing tools earlier this year after reports surfaced that there were numerous errors in these articles.
Major publications such as the New York Times and Bloomberg have blocked OpenAI web crawler over fears the AI firm is using their content to train ChatGPT. In April, Insider's global editor-in-chief Nicholas Carlson said the company would start using AI to generate outlines for stories and craft headlines, for example.