The Quartz Intelligence Newsroom, a generative AI writer, has churned out 20 stories today so far. To do this it digests the news from a wide range of sources, including local news affiliates, Reuters, and NPR broadcasts, before regurgitating it back in a flaccid form. Some of its sources are themselves generated using AI. So AI slop is devouring AI slop to produce AI slop. It’s a wonderful information ecosystem we’re building here folks.
As first reported by Aftermath, Quartz is one of the last remaining jewels in the G/O Media empire. The business-oriented website once published work from journalists who’d come from The New York Times, the Wall Street Journal, and Bloomberg. Now it scans the work those publications do, cobbles them together into something resembling news with a machine, and vomits it onto the internet.
G/O (which owned Gizmodo from 2019 to June 2024) has been big on using AI to do less with more. Journalists cost money. We like to eat and pay rent and G/O’s money men would rather not pay for that. So, instead, they’ve turned on various robots they assume can do what we do but faster and cheaper.
For months the Quartz Intelligence Newsroom was relegated to churning out ChatGPT-style summaries of earnings reports, but it’s recently gotten into the op-ed and reported news game. On January 23, it published its first big hit. It turned two articles about something Larry Fink said at Davos into a news hit. It followed that up by writing the same story about Samsung phones twice under two different headlines.
To its very small credit, Quartz’s AI news writing machine doesn’t pretend to be anything other than what it is. “Quartz Intelligence Newsroom uses generative artificial intelligence to report on business trends. This is the first phase of an experimental new version of reporting,” it says under every story. “While we strive for accuracy and timeliness, due to the experimental nature of this technology we cannot guarantee that we’ll always be successful in that regard.”
It also provides links to the sources it scrapped at the top of every story. It’s a list of stuff you should read instead of the Quartz article. Earlier today it published a profile of Liang Wenfeng, the founder of DeepSeek. Its sources were a BBC and Forbes writeup of the LLM, an article from Italian news site FIRSTOnline, and an article from the site Devdiscourse.
As Futurism first pointed out, Devdiscourse is, itself, an AI-generated content farm. It looks like a real website at first blush but the byline “Devdiscourse News Desk” turns out a startling number of stories seemingly cobbled together from other news sources and thrown under AI-generated images. Under the headline “Colombian Migrants Return Amid Tensions with US,” a smiling couple walks down the beach holding suitcases. The man is dressed in a shirt that appears adorned with the AI’s approximation of the Colombian flag. It’s using the wrong colors.
“This image is AI-generated and does not depict any real-life event or location. It is a fictional representation created for illustrative purposes only,” the site explains. Thanks, I wouldn’t have noticed otherwise.
Quartz’s AI newsroom is scraping legitimate news sources, mixing them with dubious AI slurry, and then generating its own AI slop.
+ There are no comments
Add yours