ST. LOUIS — An artificial croak from a thought-to-be-dead alternative news outlet caused many to cry foul across St. Louis.
For years, articles from the Riverfront Times drew interest, outrage, and laughs. But people especially took notice when an article was posted out of nowhere a month after an unnamed entity bought the periodical and subsequently laid off its entire staff.
Many accused the article of being written by artificial intelligence.
PREVIOUS COVERAGE: Riverfront Times sold, newspaper's editorial staff laid off
The article reported on Missouri Governor Mike Parson's decision to deny clemency of death row inmate David Hosier. Some nuances of the story caught the attention of the region's journalists, including:
- The dry, personality-free nature of the article's writing
- The article's author, Ryan Smith, never wrote an article for the website before this story and didn't have any social media links to personal or professional accounts
- The story was posted days after the news originally broke, and a day after the state executed Hosier
"You can judge for yourself, but having used ChatGPT to produce fake news stories about MO pols competing in ladder matches, this looks like AI," St. Louis Public Radio Political Corespondent Jason Rosenbaum said of the article on X, formerly known as Twitter.
Rosenbaum previously posted on X that the only joy he's gotten from AI was asking ChatGPT to write fictional news stories about all eight Missouri Secretary of State candidates competing in professional wrestling matches.
RFT deleted the article two days after backlash mounted. A screenshot of the article can be seen here:
5 On Your Side ran the article through numerous free AI detectors, including ZeroGPT, Copyleaks, Merlin, Leap and Undetectable AI. Each of the checkers reported there was a 100% likelihood the article was written by AI and had 0% human input.
To double-check our work, we reached out to Webster University Communications Instructor and media literacy expert Julie Smith who specializes in media literacy and combatting disinformation. Smith ran the article through three other AI detectors, WinstonAI, GPTZero, and ContentAtScale, each of which came to the same conclusion: the text was AI generated with a 100% likelihood.
"I don't typically trust [AI Detection Tools] 100%, but if they ALL say the same thing, that gives them some credit," Smith said.
5 On Your Side attempted to contact the Riverfront Times for confirmation of the article's artificial intelligence author, but each of the emails in the publication's "Contact Us" section and Smith's RFT email bounced back.
It's unclear whether RFT will lean on artificial intelligence for future articles, but Smith said readers should be skeptical of articles they suspect were written by AI, especially during an election year.
"AI-generated stories can say anything, and people will believe them because we scroll so quickly through frictionless sites - very few people will stop to see if anything is AI-generated (especially if they like the story)," Smith said. "Journalists have standards, editors and put their names on stories. They've also been trained."
Smith shared multiple tips for how to spot AI-generated stories, including:
- Look for emotionless writing without personality
- Click on the reporter's name and see whether they have other stories or are on social media. If a photo is attached to the reporter's bio, run it through TinEye or Google Reverse Image Search to see what other sites use that photo
- Do "lateral reading" by checking other sources to see if anyone else is reporting on this story
Readers can also ask how the message is designed to get their attention, what information is left out of the reporting, and who benefits or profits from the article.
Top St. Louis headlines
Get the latest news and details throughout the St. Louis area from 5 On Your Side broadcasts here.
