Is it safe to use AI for my business? What never to paste, and how to spot a scam

It can be safe for everyday work if you keep private details out and check your data settings. In our view, the bigger risk for most owners is scams, not the tool. That includes scams dressed up as AI.

Why it matters

One pasted card number or login can cost far more than any time AI saves. Customers trust you with their details. And your email is the key to your other accounts. A few free habits from the NextGen Business Community protect all of this.

Do it with AI: scrub before you paste

Before a note goes into AI, swap out the private parts. Practise on a made-up note first.

Below is a note I want to use with AI later. Help me scrub it first.
1. Replace every person's name with [NAME 1], [NAME 2] and so on.
2. Replace phone numbers, email addresses and street addresses with [PHONE], [EMAIL] and [ADDRESS].
3. Replace any account, card or invoice numbers with [NUMBER].
4. Keep everything else the same.
5. Under the clean note, list what you replaced.

The note:
[A MADE-UP PRACTICE NOTE, OR A NOTE YOU HAVE ALREADY CLEANED UP BY HAND]

Sample (made up): a practice note from Hillside Auto Repair. It has a customer's full name, cell number, street address and car. It also says the brakes get done Thursday.

What a good answer looks like: the name, number and street become [NAME 1], [PHONE] and [ADDRESS]. The car and brake job stay. A list at the bottom shows the three swaps.

Scrub real notes by hand before you paste. AI can double-check one you've already cleaned. Never paste card numbers or logins, not even to have them removed.

Tools checked 2026-09-26: you can use ChatGPT's free plan for this. Claude or Gemini work too.

What to watch for

Keep private things private. Don't paste card numbers, bank logins, Social Security numbers, passwords or health details into any AI tool. Before you paste customer messages or lists, swap full names for first names or initials and remove phone numbers and addresses. Check your AI tool's data settings (the next line shows where in ChatGPT, as of 2026-09-26).

In ChatGPT, it's under Settings, then Data controls: the switch called "Improve the model for everyone." Turn it off if you don't want your chats used to train OpenAI's AI (checked 2026-09-26).

Signs of a scam. Warning signs from the Federal Trade Commission (FTC):

They pretend to be someone you know: a government agency, your bank, a utility or a supplier. Caller ID can be faked.
There's a problem or a prize. You owe money, or you've won something.
They rush you so you can't stop and check.
They want paying in an odd way: gift cards, wire transfers, cryptocurrency or a payment app.
They ask for the code your bank or email just texted you. Anyone who asks for it is a scammer.
A bill you don't recognize, for supplies, ads or a directory listing nobody ordered.

When in doubt, don't click. Contact the company with a number or website you already know.

Strong passwords. CISA, the federal cyber safety agency, has three password tips. Make each one long: at least 16 characters. Make it random. And use a different one for every account. Four to seven unrelated words work well. A password manager can store them. Also turn on two-step login for email, bank and social media.

See it done safely. The AI Starter kit has five ready-made AI routines for everyday jobs. Routine 2, "Customer message → reply draft in your voice", builds these privacy habits in. The kit is $47, one time. Not a subscription. It's a simple first step with AI, from NextGen.

Get the AI Starter kit

Common questions

What should I never type into ChatGPT?

Card numbers, bank logins, Social Security numbers, passwords and health details. Also take customer contact details out.

Does ChatGPT keep what I type?

Yes, unless you delete them. Chats you save stay in your history until you delete them. Deleted chats are wiped from OpenAI's systems within 30 days, with a few safety and legal exceptions. On a free or Plus account, OpenAI may use your chats to train its AI. That is on unless you turn it off. To turn it off, open Settings, then Data controls. Switch off "Improve the model for everyone." This covers new chats from then on. If you tap thumbs up or down on a reply, that whole chat may still be used. Want a chat that is not saved or used to train? Pick Temporary chat. OpenAI may keep a copy for up to 30 days for safety.

How do I spot a "make money with AI" scam?

Watch for big promises, a secret "proven" system, reviews you can't check, and a cheap start that turns into big bills. If they rush you or want gift cards or crypto, walk away.

How do I make a strong password?

Long, random and one per account, plus two-step login.