Hoodline is Using AI Generated Writers to Make AI Generated Articles

Hoodline’s most prolific writer. (Photo of Pepper the robot by Alex Knight via Wikimedia commons)

In a case of “I hate to say I told you so,” let me just say…I absolutely told you so.

Back in 2019, almost exactly five years ago actually, I wrote an article for the SF Examiner called “Can robots do journalism?” In the article I called out Hoodline for using big data and artificial intelligence to write listicles like “The 5 Most Popular Movies in San Francisco this Weekend” or “Where to Eat Now in Oakland”. Basically they were taking data from places like Fandango and Yelp and having AI plug the info into templates and turning those into articles. Basically Madlibs for robots.

Sure it was innocuous enough, and honestly, as far as AI writing goes, it’s a decent way to use the technology and not be too gross. Like, if this is info your audience really wants, having AI write these dum-dum little articles can free up your writers to do real journalism.

The thing that really pissed me off though, was that the company had raised $10 million to try and solve the problem of disappearing local news sources. But they had more data scientists than writers and editors, and weren’t actually trying to fix the main issue: that the news business is broken and it’s harder than ever to fund journalism.

To quote myself (I love getting to say that):

“Hoodline claims it’s trying to fill the hole created by the disappearance of local publications. But nobody gives you $10 million just so you can find a quicker way to tell people which Marvel movie is at the box office…That $10 million should have been used to hire brilliant minds to tackle the question of ‘How do we fund good journalism in the 21st century’. Instead it’s being used to figure out a way to one day get rid of journalists altogether and ultimately make the millionaire investors even richer than they already are.”

And apparently, that “one day” is upon us…or at least a really shitty version of it is.

Let me pause for a brief, but very opportune moment, to ask you to join the BAS Patreon because we literally need your help so we can pay people to write things.

On Tuesday, Joshua Bote broke a story for a new publication called Gazetteer that could be summed up by the sentence “Stuart absolutely fucking called it!” That sentence isn’t actually in the article, but the piece is well reported, and goes into deep detail showing how Hoodline has been using AI to generate news stories and journalist profiles.

Nina Singh-Hudson’s Hoodline byline from May 9, 2023. This is not a real person despite there even being a bio for them. This screenshot was originally found by Joshua Bote at Gazetteer.

That’s right, they have fake writers writing really poorly written articles, and they haven’t been disclosing any of it. And here’s the kicker: it’s been going on since May 2023!

As Bote writes “Up until late last year, too, Ng, Vargas and Singh-Hudson had human-sounding biographies. They were all removed sometime late last year as part of a website refresh, prior to the April update of the site.” So Hoodline went as far as making up fake biographies of their fake writers.

Now I know some of you are saying “Hey Stuart, I’ve seen you use AI generated images on BrokeAssStuart.com before, how is this any different?” And that’s a totally fair question.

First and foremost we always disclose when we use AI. It’s the ethical thing to do. Secondly, we don’t have any fucking money. As it is, Broke-Ass Stuart costs over $10,000 a month to run and I’m amazed every single month we’re still around. Since we often can’t afford to license art or pay people to create original artwork for us, we use Creative Commons materials as much as we can. But sometimes, there just isn’t anything out there that fits the bill, so we’ll use AI images to fill in the gap. It’s certainly not ideal, but I feel better about it knowing that we give hundreds of artists thousands of dollars worth of free publicity every year. That said, if every single person who saw this article joined the BAS Patreon, we might just be able to actually license more original art (nudge, nudge, wink, wink).

While you should really read the article that Joshua Bote wrote to see his great reporting, here are a few things I must note:

– Zack Chen, CEO of Hoodline parent company Impress3, published a letter admitting that Hoodline has been using AI as described, and justifies it by saying it was necessary so they could “build the traffic and revenue needed to employ more experienced, full-time journalists.”

– Chen’s letter noted that a little badge will now appear next to AI generated content.

– It worth noting that Hoodline still has one real writer, Steven Bracco, and his articles are the only ones that seem to get any real traction.

– Hoodline was originally founded in 2010 as Haighteration, a blog focused on the Lower Haight. In 2014 it relaunched as Hoodline and had journalists coving their own neighborhoods. Since then it’s been bought a bunch of times for various large sums of money.

– The current ownership is not the same folks who owned it when I wrote the article in 2019. But they are owned by a company that owns a number of other properties, in a multitude of fields. They are most likely not broke as fuck.

– Update: Zack Chen reached out to let me know know that “Per your questions about the Use of AI page, the information there was largely on our Terms of Service page prior, but we wanted to increase transparency by making those notes more prominent and directly in the footer of each page.”

For the rest of the story though, I’ll let you read the Gazetteer article right here.