Get the newsletter

Home › Guides › Staying safe with AI: privacy, scams, deepfakes and checking answers

Guide

Staying safe with AI: privacy, scams, deepfakes and checking answers

Simple habits to protect your privacy, spot AI-powered scams and deepfakes, and double-check what AI tools tell you.

5 min read · Updated · Written and checked by real people

AI tools can be wonderfully helpful. Like any powerful technology, though, they come with a few risks. The good news: a handful of simple habits will keep you, your family and your information safe. You don’t need to be a tech expert, just a little cautious.

1. Protect your privacy

When you type something into an AI chat tool, that information is sent to the company that runs it. Depending on the service and your settings, conversations may be stored, reviewed or used to improve the AI.

Simple habits:

  • Don’t share what you wouldn’t post publicly. Avoid passwords, Social Security numbers, bank details, medical records or private information about other people.
  • Check the settings. Most major assistants, including ChatGPT, Gemini and Claude, have privacy or data controls. Look for options about chat history and whether your conversations are used for training.
  • Use work-approved tools for work. If your employer provides an AI tool, use that one for work documents. It often has stronger privacy protections.
  • Be careful with uploads. Before you upload a document or photo, ask: “Would I be okay if someone else saw this?”
  • Remove personal details first. You can often get the same help by replacing names and numbers with placeholders like “[Client A]”.

2. Recognize AI-powered scams

Scammers have always tried to trick people. AI simply helps them do it faster and more convincingly. The FBI warns that criminals now use AI to write convincing messages, create fake profile photos, clone voices and even produce fake video calls (FBI IC3).

Voice-cloning “emergency” calls

One common trick: you get a frantic call from someone who sounds exactly like your grandchild, child or friend, saying they’re in trouble and need money right away. The U.S. Federal Trade Commission explains that a scammer needs only a short clip of someone’s voice, perhaps from a video posted online, to clone it (FTC).

What to do:

  • Don’t trust the voice. Hang up and call the person back on a number you know is theirs.
  • Create a family “safe word.” The FBI recommends agreeing on a secret word or phrase with your family to confirm identities (FBI IC3).
  • Watch for pressure and unusual payment requests. Requests to wire money, send cryptocurrency or buy gift cards are major warning signs (FTC).

Phishing emails and texts

AI makes scam messages more polished, so spelling mistakes are no longer a reliable clue. Instead, look for:

  • Urgency (“Your account will be closed today!”)
  • Requests for passwords, codes or payment
  • Links or attachments you weren’t expecting

When in doubt, go directly to the company’s website or app yourself rather than clicking a link.

Fake investment and “AI trading” offers

Be very wary of anyone promising guaranteed returns using “AI.” The FBI notes criminals use AI-generated videos and fake websites to promote fraudulent investment schemes (FBI IC3).

3. Spot deepfakes

A deepfake is a fake image, video or audio clip made with AI to look or sound like a real person. They’re getting harder to spot, but the FBI suggests looking for subtle signs (FBI IC3):

  • Distorted hands or feet, odd teeth or eyes
  • Blurry or irregular faces
  • Strange shadows, accessories or watermarks
  • Lag, unnatural movements or a voice that doesn’t quite match

Even better than spotting flaws: check the source.

  • Who posted it? Is it from a trusted news outlet or an official account?
  • Can you find the same story reported elsewhere?
  • Some images now carry “Content Credentials,” a kind of digital label showing how they were made. You can upload an image to the free Content Credentials Verify tool to check.

Protect yourself, too: consider making social media accounts private and limiting how much video and audio of yourself you share publicly (FBI IC3).

4. Double-check what AI tells you

AI chat tools can sound very confident even when they’re wrong. They sometimes invent facts, quotes, statistics or even sources. This is often called a “hallucination.”

A simple checking routine:

  1. Ask for sources, then click them. Make sure the source exists and actually says what the AI claims.
  2. Cross-check important facts with a trusted website, such as a government agency, a well-known news outlet or an official organization.
  3. Be extra careful with health, legal and money questions. Use AI to prepare questions, then talk to a qualified professional.
  4. Check numbers and dates yourself, especially in anything you’ll share or sign.
  5. Notice when something seems “too neat.” Real life is messy; perfectly tidy answers deserve a second look.

5. Keep your accounts secure

  • Use strong, unique passwords (a password manager helps).
  • Turn on multi-factor authentication wherever you can.
  • Keep your phone and computer updated.

The U.S. Cybersecurity and Infrastructure Security Agency recommends these same basics for everyone (CISA).

If something goes wrong

  • Report scams to the FTC at ReportFraud.ftc.gov and online fraud to the FBI at ic3.gov.
  • Contact your bank right away if you sent money or shared account details.
  • Change passwords for any affected accounts.
  • Tell someone. Scammers count on embarrassment. Sharing your experience helps protect others.

The takeaway

Staying safe with AI comes down to a few calm habits: share less, pause before acting on urgent requests, verify through a trusted channel, and double-check important answers. You’re in control, and a moment of caution goes a long way.

Sources

More guides