How should we feel when an AI journalist operation falsely accuses someone of murder? How about if the person falsely accused is a public figure? I know there are some questions lately about where liability should land when an AI system hallucinates, but here it seems pretty clear who is at fault: a nonsense peddling AI sludge factory called “Hoodline” run by a company called Impress3.

We’ve written multiple times about the growth of absolutely horrible AI sludge journalism, in which crappy (often legacy) news sites are replacing reporters with terribly written, prone-to-lying, AI journalists, with apparently zero editorial review. As I’ve said, I do think there are places where generative AI tools can be useful in journalism, but it’s not in writing stories independently without any review. Indeed, I use AI tools to interrogate everything I write (including this article, which the AI didn’t much like) before I then have a human editor review it.

But apparently, some others don’t much care, including “Hoodline.” We actually mentioned them a few months ago in one of our stories about AI sludge news sites.

A few days ago, I was reading the news recommended by Google News, and a story caught my eye, claiming that the San Mateo County DA had been charged with murder!

This seemed like it would be really big news! Especially in San Mateo County where I live. And I hadn’t even heard anything about it. The article starts out by claiming again that the San Mateo DA had been charged with murder.

Following a preliminary hearing, San Mateo County District Attorney John Caisiano Thompson, age 56, is being held to answer on all charges related to the alleged murder of a woman in his residence despite the victim’s body remaining unfound. According to an official X account from the San Mateo County District Attorney’s Office, Thompson was apprehended after a collaborative investigation by the DA’s Office and the East Palo Alto Police Department, which initially began as a search for a missing person.

When I first read this, I wondered if the San Mateo County DA’s office had handed this off to another DA to avoid the conflict of interest. But there was nothing of the sort discussed in the article. Indeed, the more I read through the article, the more it became clear that this wasn’t actually the San Mateo County DA at all. It was just some guy they arrested a while back finally being charged.

Yes, there’s a little tiny “AI” next to the byline of the reporter, which should have clued me in. But I wasn’t even looking at the byline, and even if I had, I probably wouldn’t have noticed the little “AI” icon.

So, what happened? Well, the Nafnlaus account on Bluesky figured it out.

Haha, I can tell exactly what happened here!Left: The original post.Right: Rendered as plaintext for the AI to parse

(Enn) Nafnlaus (@nafnlaus.bsky.social) 2024-10-09T13:46:25.845Z

It was, in fact, trying to make a story out of a tweet by the San Mateo County DA’s ExTwitter account. But when the AI parsed it, it merged the name of the account “San Mateo County District Attorney” with the start of the tweet, which begins with the guy being charged. So merged together, it looks like that name is the DA.

When parsed, looks like:

And the AI is too stupid to notice this or to do a simple look-up to realize the San Mateo County DA is actually Stephen Wagstaffe. You know, the kind of thing a human might actually check.

Anyway, this is yet another reason why this is not the right way to be using AI in journalism. It makes me wonder why Google is featuring a site like Hoodline in Google News when it publishes this kind of nonsense. This should also alert people that any site run by Impress3 is not at all trustworthy, especially Hoodline.

Leave a Reply