Google has a new AI tool for journalism—but should we trust tech to meddle in newsrooms?

Google has a new plan to help save journalism—or at least make it easier for journalists to work. The company has been pitching a generative AI-powered product, nicknamed Genesis, to major publishers including The Washington Post, Wall Street Journal owner News Corp, and The New York Times—the latter of which reported on those pitches earlier this week.

A kind of specialized ChatGPT for the media industry, Genesis can be fed information about current news events and produce written journalistic stories as an output. One executive who was shown the tool told The New York Times they found it unsettling.

Google seems to think its new product amounts to a quick fix to save journalists time and effort. As is the case across all industries, the use of AI is meant to make the production of journalism more efficient. Extra efficiency is crucial in a media market where mass layoffs have become the norm, and those who remain are often expected to do more with less.

More than 17,000 people were fired from media jobs in the first half of 2023, according to a new report from outplacement firm Challenger, Gray & Christmas—a record number for the first six months of any year. In theory, AI could help fill gaps in staffing brought about by those cuts. 

But it’s worth remembering that the media industry is in such a dire state right now precisely because of big tech’s moves. Less than a decade ago, large swathes of the media took part in the great “pivot to video,” egged on by companies like Facebook (now Meta) that were eager to populate their platforms with ever-new forms of content. Staff producing text copy were laid off in favor of multimedia specialists who could perform on camera—only for tech platforms to soon pull the plug on the video experiment, an about-face that left media outlets who had refigured their entire business model high and dry.

“I think it’s quite obvious that Google and Meta don’t do things out of the goodness of their hearts,” says Paul Wiltshire, a senior lecturer in journalism at the University of Gloucestershire. “The [media] industry has had its fingers burnt by being too reliant on tech, and taking perhaps more seriously than we should have promises from Google and Meta that they do want to support and invest in journalism.”

To be sure, newsrooms have benefited from some of its collaboration with the tech sector, such as the ongoing Google News Initiative, which backs journalism projects around the world and provides training for companies and colleges to keep skills up to date. And some would argue it’s important that the media industry gets a handle on AI early on and not repeat the flat-footed mistakes of the early web, when news publishers chose to ignore or downplay the risks and potential of the new technology—with catastrophic results.

“The reality is big tech isn’t going anywhere and they provide the platforms and tools for audiences to engage with news, in whatever form that takes,” says Sarah Schijen, a media strategist who previously worked at Vogue and BBC News. She believes it’s important that the media engages with AI before it’s too late. “In order to keep providing truthful, impartial, reliable news, it’s necessary to engage.”

But at what cost does engagement come? Some newsrooms, including at G/O Media, are posting entirely AI-generated stories to their website without any human oversight—many of which are riddled with errors. And we know AI can perpetuate biases that are already inherent in media coverage, including racism and sexism based on the data it is trained on. 

That’s not to say Genesis would necessarily fall victim to the same trappings as its AI peers, but it does suggest newsrooms would be wise to tread with caution. “We need to have a healthy skepticism about this.,” says Wiltshire. “Looking at what this Genesis idea is so far, you could be forgiven for thinking there’s quite a lot of smoke and mirrors here.”