Hi! Happy Amazon Prime Day(s) to all who celebrate. May your week be filled with discounted home goods and impulse gadget buys. Today’s Useful Thing: how to tell if a deal is actually a deal.

But first, I fell for an AI spearphishing scam. Well, sort of. Then AI helped me spot it. Here’s how large language models are making scams much harder to detect—and what you can do about it. Plus, another week, another pair of smart glasses. This time: more stylish, more Kylie Jenner.

🔔 Reminder: Today’s full newsletter—with all the tips, advice and recommendations—is for paid subscribers. You can sign up here to read the full edition. 🔔

Joanna Stern and the email scammer who pretended to be author Mitch Albom. Photo illustration: Getty Images, New Things

Obviously I should have known: Mitch Albom has better things to do than email me. 

But I grew up reading Tuesdays with Morrie and The Five People You Meet in Heaven, so when famed author “Mitch Albom” popped up in my inbox to say he enjoyed my new book—that it had “stayed with me longer than most”—I was excited. So excited I texted my mom immediately. And my wife. 

Now, to be clear, my spidey senses went off. The email address wasn’t from Mitch’s website domain. It was close, but not quite. Plus, the email included the line, “I’m not writing to pitch anything,” which, in my experience, is something people say right before they pitch you something.

So I wrote “Mitch” back and asked for more proof. A Mitch Albom CAPTCHA, I joked.

“Mitch” replied with a compelling answer, full of details about his life and books. But most of that could have come from the internet. So I asked for a voice memo or audio recording about his favorite section of my book. 

“Mitch” said he couldn’t do that. By then, I knew “Mitch” wasn’t the real Mitch. I even confirmed it with HarperCollins, the publisher of both my book and many of Mitch’s—the real Mitch, not air-quotes Mitch.

But I kept the conversation going to figure out why “he” had gone to such lengths to spearphish (a highly targeted phishing attempt) me. Around email eight, “Mitch” finally explained. “The honest version: I started this with a loose intention of eventually connecting you with someone I work with, a book marketer who’s exceptional at getting the right books in front of the right audiences.”

Ah ha! For weeks, I’d been hearing from book-marketer spammers. Finally, one got me with a smart impersonation and clean, straightforward writing. I asked “Mitch” if he was using AI to write this. “Honestly? No chatbot. No LLM,” he wrote back. “Just me, a keyboard, and apparently way too much time to craft emails to a tech reporter who was determined to catch me out.” Sounds like AI to me! 

I share this not only to risk my reputation as an eagle-eyed tech journalist, but because this is exactly how AI-era scams work now. Broken English and Nigerian princes have been subbed for believable flattery, crisp prose and a lot of personal detail.

Here’s what I learned about how to protect myself in the future:  

This newsletter was written and curated by Joanna Stern and Adele Lowitz. 

Keep Reading