Aspen Policy Academy

Think You’ve Been Scammed? AI Is Making It Harder to Tell – Here’s What to Do

  • Article Published August 8, 2025

This article originally appeared in The AI Journal on August 8, 2025.

By Betsy Cooper

I came down from my home office to hear our family’s nanny, “Trina,” on the phone in distress. “I think I’ve been scammed online,” Trina told me. “What should I do?” 

Trina explained that she had lost her social security card and was looking to get a new one. Like many of us, she had Googled the options and clicked on the first link, a website called SSNSimple.

That company had collected her personal information (including her name, address, and social security number), charged her $39, and then claimed it had filed a request on her behalf. Trina had then received a letter back in the mail saying that her application could not be processed. She then called the local Social Security Administration, who said they had no record of the transaction. Trina then learned that you can request a social security card for free from the Social Security Administration. No fee was required in the first place. 

Trina was distraught. What were the risks to her personal identity? What should she do next? And I quickly joined in her anxiety. Here I was, a so-called cybersecurity professional. And while I’ve spent years advising people on simple steps that they could take to protect themselves before problems arise, I didn’t have a similar list for what to do after a scam had occurred.

Browse Related Articles

Green code against a black screen.

In Pentagon-Anthropic standoff, AI is real-time testing the balance of power in future of warfare

This article originally appeared on CNBC on February 27, 2026.
[Image] Two women sitting across from each other, leaning over a table.

Apply Now for the 2026 Building Nonprofit Advocacy Capacity Short Course

Learn how to level-up your nonprofit’s impact.
Mock code for an AI Large Language Model (LLM) that could intelligently answer questions.

How to Manage Misinformation in Large Language Models

This article originally appeared on Tech Policy Press on February 25, 2026.