Photo-illustration by Newsweek/Courtesy of Matthew Albracht

Matthew Albracht

By Matthew Albracht

0

Share

Newsweek is a Trust Project member

See more of our trusted coverage when you search.Prefer Newsweek on Googleto see more of our trusted coverage when you search.


Over the last few weeks, I found myself pulled into what appears to be a new kind of scam—and one that feels like a troubling sign of where things are headed. Not a phishing email from a fake bank. Not a prince needing help moving funds. This was something more sophisticated: an AI-generated scam, built around the identity of a real person and carried out through a polished, surprisingly intelligent email exchange.

And for a while, I fell for it.

Thankfully, not all the way. I didn't hire anyone, send money or hand over anything sensitive. But I did spend real time, real thought and real energy in correspondence that, in retrospect, clearly was not what it claimed to be. What unsettles me now is not just that someone tried to deceive me. It's how high-level the deception was, and how soon AI-generated scams may become commonplace.

Read More on My Turn

I Noticed a Rash After Using a Hot Tub, Then Doctors Revealed I Had HIV 6 min read

I Noticed a Rash After Using a Hot Tub, Then Doctors Revealed I Had HIV

My 12-Year-Old Daughter Woke Up As Usual—That Same Day Life Changed Forever7 min read

My 12-Year-Old Daughter Woke Up As Usual—That Same Day Life Changed Forever

I Loved Tanning. 22 Years Later, I’m still Living With the Consequences9 min read

I Loved Tanning. 22 Years Later, I’m still Living With the Consequences

Matthew Albracht nearly fell for a new AI-generated scam.

...

About six years ago, I wrote and self-published a book. After about a year of promoting it, I closed that chapter of my life and fully moved on. So when a woman calling herself Becky reached out praising the book and offering continued promotion, my ego felt a bit flattered. But I was uninterested. Like anyone who has spent years running a nonprofit, as I have—or really any organization—I am used to cold pitches. I ignore pretty much all of them.

Initially, I planned to ignore this too.

But timing is part of how these things work. These days I'm working on a new book proposal and, more urgently, a collaborative, multi-author Substack-based media platform exploring social and cultural issues through a psychological and systems-informed lens. The day before “Becky” reached out, I had been thinking about how I’d eventually need marketing and strategic support, especially for launching that platform—though not quite yet. So when her email landed, the timing felt annoyingly synchronistic. I thought, fine, maybe I’ll lightly engage and see if there’s anything useful here.

I looked into her background. I found a Reedsy profile. The person whose identity seemed attached to the outreach had real credentials and a professional track record. Nothing immediately screamed scam.

Then the conversation deepened—and this is the part that still gets me.

The messages were not just polished. They were smart, with the feel of someone masterful at her craft. "Becky" was not merely flattering me (if anything, I would have welcomed a bit more of that). She was asking the kind of questions and offering the kind of strategic insight a real collaborator might. She reflected back that what I was building was bigger than a single book—more like an ecosystem in which media, policy and community initiatives might all feed each other—and she accurately described the work as integrating "intrapersonal healing, relational repair, and structural change." All accurate to my deepest intent, and not the sort of complexity you expect from a scam, at least not the scams of old.

This was not crude flattery. It mirrored the actual depth and architecture of what I'm trying to build at a level that made me think: This person is quite perceptive. Maybe I've stumbled onto someone I could really use. A real partner, I let myself dream.

So I continued the dialogue. At some length. Not because I was about to hire someone cold off the internet, but because the exchange itself was drawing useful thinking out of me. My excitement grew. I started to feel that energizing sense that perhaps all this might actually come together—that the “universe” had sent me, strangely, an intuitive strategist-publicist.

Reader, the universe had not.

Eventually "Becky" sent over a proposal with pricing tiers. I replied that before I would ever consider hiring someone, I would need to speak live. Voice to voice. I wanted to hear how she thought in real time and verify that she was, in fact, an actual person with skills, not just someone using AI to help craft responses to me, something I had started to wonder about.

That is when the whole thing cracked open.

I received a bizarre email that seemed to reveal a glitch in the machinery. The message came from her account, but it was written as though it had been composed by me—language I never wrote, sent from her address, turning the whole exchange inside out. The mask had slipped. Whatever was behind this had malfunctioned. And dishearteningly, I realized: It had never been real. My seemingly heaven-sent strategic partner was not to be.

I cannot prove with certainty exactly what was happening—some hybrid of spoofing, automation and possibly human supervision. But whatever the precise setup, the effect was the same: a deceptive interaction that used strategic mirroring, polished language and high-level feedback to generate trust.

As I later learned, this wasn’t an isolated incident. The Federal Trade Commission has warned that scammers are increasingly using AI to enhance impersonation schemes, and the FBI has issued alerts about AI-assisted fraud involving both text and voice messages. In the publishing world, organizations like Writer Beware, along with reporting from The Guardian, have recently documented a rise in AI-driven scams targeting authors.

We tend to think of scams as laughably obvious. But this met me right where I live: in ideas, in ambition, in the vulnerable and hopeful terrain of building something meaningful. It did not merely flatter me. It appeared to understand me. That is a much more powerful hook.

And it feels like a preview. If this kind of AI-assisted scam is already here in written form, what comes next isn’t hard to imagine: real-time phone calls that sound convincingly human and can carry on fluid conversation, video that looks real enough, entire exchanges tailored to your professional aspirations, scraped from your public work and fed back to you with warmth and just enough insight to keep you leaning in. Add bad actors, stolen identities and financial motive, and things get strange fast.

In my case, the scam failed because I insisted on one old-fashioned thing: an actual conversation before any money changed hands. That may still be enough, for now. But we are rapidly entering a world in which our ordinary cues for credibility are breaking down. A polished email, a real-looking profile, a surprisingly insightful message—none of these may be reliable evidence that you are dealing with who you think you are. And soon, as the technology progresses, even a live “voice” on the phone or a “face” on a video call may not be either.

Trust used to require discernment. Now it may require verification of a kind that could soon become much harder to come by, a disturbing possibility most of us haven't yet begun to reckon with.

And if you ever receive an email from someone who seems uncannily good at understanding your deepest work, your long-range vision, and the exact architecture of what you are trying to build, pause before feeling too flattered.

You may not be talking to your next strategic partner.

You may be talking to a machine dressed as one.

Matthew Albracht is a writer, advocate and Board Chair of The Peace Alliance. He writes about politics, culture and collective healing on his Substack.

All views expressed in this article are the author’s alone.

Do you have a personal essay you want to share with Newsweek? Send your story to MyTurn@newsweek.com.

View 0 Comments

Request Reprint & Licensing

Submit Correction

View Editorial & AI Guidelines