The Chicago Sun-Times put out a summer reading list for its readers, a standard practice for publications that still sport an arts and culture section in its pages. The only problem? Several of the titles were made up. The paper used AI to generate a list of largely fake books. So, if you happened to peruse the list to see what you might like to read over your summer vacation, you’ll have to consult another publication.

The list clearly wasn’t vetted before it was published, and it was even featured in a physical issue of the journal. They won’t be able to subtly hit “delete” on this one. Gizmodo picked up the story and reports,

The list was reportedly submitted for publication by freelance writer Marco Buscaglia, who spoke with 404 Media. Buscaglia told the news outlet he was “completely embarrassed” and that while he does sometimes use AI to create content, he typically checks it before submitting it. Buscaglia didn’t immediately respond to Gizmodo on Tuesday, including questions about what AI tool he used.

The Sun-Times didn’t take responsibility for the mishap, since a freelancer “wrote” the list, but still. If editors of a high-profile publication let something like this through their ranks all the way to the final galleys, it probably means they aren’t editing much, if at all.

Music and culture critic Ted Gioia quickly took to Substack to write about the affair with his trademark wit and biting critique of AI’s enduring problems. He writes,

AI is routinely making stupid, nonsensical mistakes that even the most incompetent employee would never make. I’ve met some incompetent journalists over the years, but none would make a boneheaded move of this magnitude.

AI is such an object of beloved hype that people are starting to forego basic methods of assuring accuracy and truth, even though the technology keeps showing its flaws. Gioia notes that tech leaders also keep telling us the technology is improving, and yet it continues to hallucinate and go haywire. When will the magic start to really work? He also mentions how, based on a survey, many people aren’t excited about AI like they tend to be about other technological developments:

This isn’t just my opinion. It’s the consensus view everywhere in the world.

A recent survey asked 48,000 people in 47 countries what they think about AI. Can you guess what they said?

People are usually excited about new technologies. Let’s fly to the moon! Let’s cure a disease! Let’s surf the worldwide web.

But AI stirs up a different reaction. People dislike it. People fear it.

I’m sure we will be seeing more of this sort of debacle in the coming months. Apart from this incident simply being an embarrassment for the Chicago Sun-Times, it also points back to the problem of AI doing a real writer’s work. In this particular case, we can see the obvious consequences. But the more journals and publishers opt for AI over the human, the more these types of mistakes will happen, and moreover, the more journalists and other writers will be out of a job. This isn’t to say human writers are infallible. They’re often quite biased and obviously make plenty of mistakes. However, AI shouldn’t replace people in the writing world. Humans write the only books worth reading, and, as we’ve now seen, they’re still the only ones who can come up with an actual list of readable books.