San Francisco Chronicle - (Sunday)
AI is helping to destroy local journalism
I got my first job in journalism in 2013 as a reporter for the hyperlocal news site Hoodline. The site had just launched by merging three small neighborhood blogs under one umbrella, and our mission was to publish stories that skated under the radar of larger publications — a mural going up on the wall of a local pizzeria or a rogue garden planted outside the DMV.
This approach proved successful at finding an audience. San Franciscans, it turned out, really wanted to know what was happening in their neighborhoods, not just in the city at large.
I found my stories by walking around my neighborhood each day. I befriended shop owners, attended neighborhood association meetings and chatted with neighbors in the dog park. Through developing valuable face-to-face relationships, my colleagues and I often landed big scoops before our larger counterparts. As our team grew over the next couple of years, so did our coverage of the city.
It was a thrilling time. Using grassroots community journalism, Hoodline managed to elbow its way into a crowded news scene. Unfortunately, it didn’t last. I quit Hoodline when it was sold to Ripple News in 2016. Nextdoor then acquired the site, and Impress3, which owns SFist, purchased it in 2020.
Now, it seems, Hoodline has started using artificial intelligence to produce most of its stories, which are aggregated from other news sites like the Chronicle. While CEO Zack Chen maintained to the Chronicle that humans still manage much of the content, there’s a catch. Leticia
Ruiz, whose byline is featured at the top of a story about the weather, isn’t a real person. Nor is Nina Singh-Hudson, who authors a lot of stories about traffic collisions, or Tony Ng, who has a penchant for police chases.
It’s particularly cringeworthy that many of these fake bylines imply the authors are people of color. Creating personas from thin air makes a mockery of the effort to improve racial diversity problems in journalism.
Seeing my old job replaced by AI is surreal, and in this context, feels particularly ironic. Oldfashioned shoe-leather reporting has been replaced by fake people who’ve never set foot in any of the neighborhoods they write about — because they don’t have feet.
But beyond the collapse of a trusted news source, Hoodline’s devolution raises larger, existential questions about news and how we receive it.
On the one hand, AI has the potential to streamline some of the most tedious and time-consuming parts of journalism. I use an AI transcription software program that saves me hundreds of hours each year; AI can crunch data and write headlines. These tools can potentially help broaden a story’s impact beyond what a small number of journalists are capable of on their own.
That said, AI is also poised to capture more of the shrinking pot of revenue that used to go to real journalists.
Nearly 75% of Americans believe local news is at least somewhat important to the health of their local communities, according to a Pew Research Center report released this week. Despite this uplifting statistic, only 15% of Americans polled say they pay for local news, yet 63% said they believe local news organizations are doing well financially. They’re not.
By the end of 2024, the United States have a third fewer newspapers than it had in 2005, according to a 2023 report from Northwestern University. Any news publications left standing will be operating with one-third of the journalists they had that same year.
The business model for the kind of intimate community journalism that so many San Franciscans dearly love collapsed even before the advent of AI. At Hoodline, all of us struggled to figure out how to make it profitable. There was no illusion that ads from local businesses alone were ever going to make it work — especially with tech giants like Google and Meta eating up most of the pie.
We need bigger, systemic solutions.
New legislation is popping up to address some of the challenges newsrooms face. The California Journalism Preservation Act would mandate that large tech companies like Google pay news organizations for profiting off their journalism. News publishers would be obligated to use at least 70% of the money to pay journalists. A second bill recently introduced by state Sen. Steve Glazer, D-Orinda, would allocate $500 million in taxes from tech giants to local journalism.
It’s an uphill battle. Google has pledged to remove links to news websites in California in protest of these bills.
But without a big shift, journalism as we know it will continue to sputter out. And it isn’t just tiny outlets like Hoodline that are in danger of going extinct or being zombified by AI.
During the past 10 years, San Francisco has lost alt weeklies SF Weekly and the Bay Guardian, and small neighborhood blogs like Mission Mission and Uptown Almanac. San Francisco Magazine no longer employs investigative journalists to produce rich, important features. Even as new publications like the San Francisco Standard pop up, sustainable financial models are elusive.
The impact on local communities cannot be overstated. Pew states that “an overwhelming majority of adults say it is at least somewhat important for journalists to understand their community’s history (85%) and to be personally engaged with their local area (81%).”
That is only possible when real human journalists embed themselves in the cities they cover. AI can never replicate that.