- Sports Illustrated published articles by fake authors with AI-generated profile pictures, Futurism reported.
- The magazine denied using AI but said it would take down the articles while it investigated.
Sports Illustrated has become the latest publication to get called out for using artificial intelligence to generate content.
The magazine said it will take down several articles after a report found they were written by fake, AI-generated authors.
Futurism identified two Sports Illustrated writers, "Drew Ortiz" and "Sora Tanaka," whose biographies appeared to be fake. Their photos are available on sites selling AI-generated headshots under the labels "neutral white young-adult male" and "joyful asian young-adult female."
The articles under their bylines are also full of the clunky language typical of AI-generated content. A piece by Ortiz's warns that volleyball "can be a little tricky to get into, especially without an actual ball to practice with."
Two anonymous sources involved in the creation of the articles in question told Futurism that Sports Illustrated had been using AI to create fake author profiles, and said that some articles under their names were generated by AI.
In a statement to Futurism, Sports Illustrated owner Arena Group denied publishing AI-generated articles but said they were removing the pieces while an internal investigation took place.
An Arena representative said the articles were "licensed content" produced by "an external, third-party company" named AdVon Commerce, in a now-defunct partnership.
In February the Sports Illustrated publisher announced it would begin experimenting with AI, with Arena CEO Ross Levinsohn saying the company would not "crank out AI content."
It's not the only digital publisher to get in hot waters over AI. Newspaper publisher Gannett paused its rollout of AI-generated sports articles in August after they were mocked on social media for bizarre phrasing. One report described a college football game as a "close encounter of the athletic kind."
Tech publication CNET, meanwhile, scrapped its own AI experiment earlier this year after articles were found to have numerous inaccuracies.
The explosion in popularity of generative AI has raised fears that tools such as ChatGPT could be used to flood the web with substandard and plagiarized content.
A report from NewsGuard in August identified dozens of sites using chatbots to churn out news stories from respected outlets, describing it as "turbocharged plagiarism."
Arena Group did not immediately respond to a request for comment from Business Insider, made outside normal working hours.