The media company, which runs several sites including Jezebel, the Onion and Gizmodo, has been publishing error-riddled stories written by AI bots, despite staff outrage over what they called “computer-generated garbage.”
G/O media bosses are doubling up on eyebrow-raising practice and plan to release additional AI-written articles soon, in accordance with internal memo obtained by Vox.
Machine-generated stories first appeared on multiple G/O sites earlier this month with titles reminiscent of “Gizmodo Bot.” One such story – a chronological list of Star Wars movies and TV shows – published on Gizmodo contained over a dozen errors, including some elements that weren’t in the right order.
Editors and reporters didn’t see the articles before they were published and were only notified that the content could be published hours before it went online, the Washington Post reports.
G/O Media’s editorial director, Merrill Brown, told staff that AI content is not any substitute for the work done by writers and editors, and acknowledged that, in accordance with the paper, “there will likely be bugs.”
As well as, he told employees that AI bots “by themselves (currently) will not be actually reliable/consistent” in one other company memo seen by Vox.
![Editorial Director Merrill Brown.](https://nypost.com/wp-content/uploads/sites/2/2023/07/EESzslQWkAAq4rJ.jpg?w=1024)
![Email sent from Merrill Brown to the G/O Media editorial team.](https://nypost.com/wp-content/uploads/sites/2/2023/07/NYPICHPDPICT000014266306.jpg?w=581)
The union representing G/O employees was quick to criticize management’s decision.
“The exertions of journalists cannot get replaced by unreliable artificial intelligence programs, known for creating lies and plagiarizing the work of real writers,” GMG Union wrote in a press release. “Our editorial offices have spent many years constructing trust with audiences – introducing computer-generated garbage undermines our ability to do our job, undermines trust in us as journalists, harms our brands and puts our jobs in danger.”
Other news outlets that attempted AI-written content — including CNET and Buzzfeed — quickly abandoned the experiment after the articles were riddled with errors and inaccuracies.
![History page for an article written by a bot.](https://nypost.com/wp-content/uploads/sites/2/2023/07/NYPICHPDPICT000014266473.jpg?w=1024)
![Several articles appeared on the pages of G/O Media, the lines of which were attributed to bots.](https://nypost.com/wp-content/uploads/sites/2/2023/07/Screen-Shot-2023-07-18-at-93110-PM.jpg?w=804)
![](https://nypost.com/wp-content/uploads/sites/2/2023/07/Screen-Shot-2023-07-18-at-92959-PM.jpg?w=796)
But Brown told Vox on Tuesday that G/O Media has no plans to stop publishing stories written and produced by bots – seemingly ignoring complaints from its employees.
“This is completely something we wish to do more of,” he said, adding that top editors will review content before publishing.
G/O Media CEO Jim Spanfeller went a step further.
“I feel it might be irresponsible to desert testing [AI]Spanfeller told Vox.
An AG/O journalist said the bot’s stories are “a disaster for employee morale.”
One other staff author said the corporate was just on the lookout for one other cost-cutting measure.
“It is a not-so-veiled attempt to interchange real journalism with machine-generated content,” a Vox employee said.