AI, writing, & book reviews

Summer Reads list with fake AI-generated book titles, published in the Chicago Sun-Times

When we launched Maverick Independent Book Reviews, we made a commitment to potential clients: No AI reviews. All books will be read by humans and all reviews will be written by humans. This to me seems like an obvious commitment to make, one that probably should go without saying, but unfortunately, not in today's world. Book reviews cover is down overall, and many places don't have any kind of book review editor. But they still want to publish reviews, or round-ups.

That's probably why, over the weekend, the Chicago Sun-Times ran a summer reading round-up that included a list of not-real-books written purportedly written by famous authors, which no real human had fact-checked before the article was published. Books like Tidewater Dreams by Isabel Allende--a title she has never written. Only five of the fifteen titles listed in the article were real. The article with the fake titles was syndicated by other newspapers including the Philadelphia Inquirer.

According to Kathleen Schmidt over on her substack, Publishing Confidental, "the list, compiled by Marco Buscaglia, was generated by AI." He's appears to take full responsibility (let's see how long that lasts) and claims to be embarrassed. Well, he should be. It's not just lazy to "write" an article like this, but he didn't even have the wherewithal to fact-check it?

Unfortunately, this is where we're going. More and more content online will be AI-generated. This includes videos. I saw one recently that purported to show Prince William opening up the door to Prince Harry and his wife Meghan and their children, with love on his face, and a big welcoming smile and hug. The video's voiceover claimed that despite differences, the brothers had all the love in the world for each other. One of the commentators pointed out the subtle signs that showed it was an AI-generated video.

But truth time? I wouldn't have been able to tell that the video was AI-generated. Not for a second. And that's the problem. We're about to be in a world of hurt--not just because human creative output is being replaced by AI-generated drivel, much of it indistinguishable from the real stuff, but also because "fake news" will now be supported by video and believed by millions around the world because--hey, there's video. What are we going to do?

If a reviewer we hire at Maverick ever does something like that, they will find themselves out of a job, we'll rescind the content and have a new reviewer write a new review, and we'll apologize for letting something slip through the cracks. But I do intend to start looking into tools to check for this.

I'm not going to spend the time to wring my hands. I'm just going to reiterate our pledge here: No AI-generated content, ever. Humans reading, humans writing, humans editing.