Got a scary text? Do this before you call back.

Today's Tip: The 30-Second Scam Check

The next time a strange text, email, or voicemail makes your stomach drop… do not call back. Do not click. Paste it into Claude first.

Let me explain…

Our neighbor Margaret, 68, retired librarian, got a text last Thursday morning. It said her Amazon account had been charged $749.32 for a laptop shipped to Nevada. There was a phone number to call "immediately" to cancel the order.

Her heart started pounding. She almost called. But she remembered something her son had shown her a month earlier… and she stopped.

Instead, she opened Claude on her laptop. She typed the entire text message in — word for word, phone number and all — and asked Claude to look at it. Thirty seconds later, she had her answer. Classic scam. The phone number had been flagged in multiple scam reports. The real Amazon never texts about charges that way.

No money lost. No call made. No stress for the rest of the week wondering if she'd been robbed.

Older adults in the United States lost roughly $7.7 billion to scams in 2024, and the schemes are getting harder to spot. AI is making the scam calls, texts, and emails more convincing than ever. But here is the quiet piece of good news… the same AI tools the crooks are using can also be used by you, in reverse, to tear their messages apart in seconds.

The trick is a single prompt you keep handy. One you can use on texts, emails, voicemail transcripts, suspicious letters, even a weird-sounding call from "your grandson." Here is the prompt that does it…

You can use this on almost anything that makes you nervous. A text saying your bank card has been locked. An email from "Medicare" asking you to confirm your number. A voicemail from a grandchild who sounds a little off and needs bail money. A letter in the mail claiming you owe back taxes.

Type what you got. Paste it in. Hit enter. Wait 30 seconds.

A good AI will walk you through what is off about the message, point out the specific red flags, and give you the correct official phone number for the real company. Your bank's real fraud line is on the back of your card. Medicare's real number is 1-800-633-4227. Social Security's real number is 1-800-772-1213. The AI will tell you these… the scammer will not.

Now, I know what you're thinking. "What if the AI gets it wrong? What if it tells me a scam is safe, or scares me about a message that was fine?"

Fair question. And it does happen — AI can be wrong. But here is the thing… you are not using it as judge and jury. You are using it as a second pair of eyes before you do something you cannot undo. If the AI says "this looks like a scam," you have lost nothing by hanging up and calling the real company at the real number. If the AI says "this looks legitimate," you still look up the official number yourself and call that one, not the one in the message. The AI never makes the final decision. You do.

Save the prompt above. Email it to yourself. Print it out and tape it next to your computer. The next scam attempt is coming — it is a matter of when, not if. Having this ready means the next time one lands, you handle it in half a minute instead of half a day of panic.

And forward this to one person today. One is enough.

★ Quick Wins

Skill Builder: Catching AI Hallucinations

Here is something nobody tells beginners… AI tools sometimes make things up. Not out of malice. Not because they are broken. They just do it. The industry word for this is "hallucination" — the AI invents a fact, a quote, a court case, a phone number, or a doctor's recommendation that sounds completely plausible and is completely false.

This matters most when the stakes are high. A made-up phone number for Medicare could put you on the line with a scammer. A fabricated side effect could worry you about a medication you take safely. A wrong tax deadline could cost you a penalty. The good news is that hallucinations have a pattern, and once you know the pattern, you can catch most of them in a few seconds.

The pattern is this: AI hallucinates most when it is asked a specific factual question with a confident tone and no source given. "What is the phone number for the Social Security office in Tampa?" is a hallucination trap. "What is the general process for contacting a local Social Security office?" is much safer, because the AI is being asked for a process it knows rather than a single number it may be guessing at.

You can also protect yourself by directly asking the AI to be honest about what it does not know. That is what the prompt below does.

One more habit worth building. When the AI gives you a phone number or a web address, do not click it from inside the chat. Type it fresh into your browser or your phone. If the AI invented it, typing it in gives you another chance to notice something is off. If it is correct, you lose about four seconds. Cheap insurance.

Think of AI as a knowledgeable friend who sometimes exaggerates. Useful most of the time. Worth listening to. But when the answer involves your money, your health, or your identity, you check it against the real source before you act. That is not being paranoid. That is being a grown-up.