G/O Media’s AI ‘innovation’ is off to a rocky start - 3 minutes read




Last week, G/O Media leadership had news for staffers at the many publications the company owns: AI-generated articles were just around the corner.

“We are both a leading technology company and an editorial organization that covers technology in world class fashion across multiple sites,” editorial director Merrill Brown wrote in an email. “So it is utterly appropriate — and in fact our responsibility — to do all we can to develop AI initiatives relatively early in the evolution of the technology.”

G/O’s early experiments with AI tools began on Wednesday through a couple of articles appearing on Gizmodo and The A.V. Club credited to the publications’ respective bots. And almost immediately, there were embarrassing mistakes.

The Gizmodo bot’s first story, “A Chronological List of Star Wars Movies & TV Shows,” contained factual errors about the in-universe chronology of the franchise, something fans were quick to point out. James Whitbrook, a deputy editor of io9, where the story appeared, tweeted that he was unaware the article would be published until shortly before. Whitbrook also said that “no one at io9 played a part in its editing or publication.” As of this writing, the original link to the story is returning an error message.

Over on The A.V. Club, a list called “The Biggest Summer Blockbusters of 2003: 10 Can’t-Miss Movies” is credited to the outlet’s bot. The article contains almost no writing or analysis, but its construction suggests that the piece is an attempt to attract cheap search traffic. The piece was also syndicated to Yahoo Entertainment.

It is unclear how the articles were assigned, generated, and if they were edited at any point by a human before going live. G/O Media didn’t immediately respond to The Verge’s questions about its editorial process and oversight of AI-written stories.

The company, which also owns publications like Deadspin, Jezebel, and The Onion, is far from the first media outlet to utilize generative AI software to produce content. From BuzzFeed to CNET, publishers have turned to AI tools to churn out material like explainer articles, quizzes, and lists, selling the pivot by saying machines would not replace human writers but instead would free up staff so they could work on more ambitious projects.

But for all of the benefits of AI tools media companies extoll, there are glaring issues — for one, the material produced is often bad or plainly inaccurate. After a litany of errors in stories produced using AI systems, human CNET staffers did the work of going back and rewriting dozens of articles. At Men’s Journal, AI-powered articles contained errors in stories about health and science that had to be corrected after publication.

Writers, editors, and other journalists at outlets experimenting with AI tools have pushed back on the encroaching technology, which at times has been deployed with little transparency — and while staffers are laid off and teams gutted. Shortly after announcing its AI initiative, BuzzFeed shuttered its Pulitzer-winning news arm. Employees at CNET recently unionized, saying workers need to have a say in how AI tools are being used at the outlet.

The Gizmodo union urged readers not to click any articles credited to the bot, calling the rollout of the articles “unethical and unacceptable.”

Disclosure: The Verge’s editorial staff is also unionized with the Writers Guild of America, East.



Source: The Verge

Powered by NewsAPI.org