Is it safe to use ChatGPT at work? What not to paste, and what to do instead
Yes, you can use ChatGPT and similar tools safely at work, as long as you treat them like any other outside service: don't hand over information you wouldn't send to a stranger, check what comes back, and agree a few simple rules with anyone else who uses them.
This guide covers the practical side for small businesses and teams. It isn't legal advice; if you handle health, financial or children's data, or work in a regulated industry, check your obligations with a qualified adviser.
What happens to what you type
When you type into an AI chat tool, your text is sent to the company that runs it and processed on its servers. Depending on the tool, your plan and your settings, conversations may be stored and may be used to improve the company's models. Business and enterprise plans usually offer stronger commitments than free consumer accounts. Read the current privacy settings and terms for the tool you use, because they change.
The safe assumption for a free account: anything you paste might be stored and seen by someone other than you.
The "never paste" list
- Passwords, PINs, security codes or login links
- Bank, card or payment details
- ID numbers (passport, national ID, tax or social security numbers)
- Customers' or staff members' personal details: names with contact details, addresses, health information
- Confidential contracts, unreleased financial results, or anything under a confidentiality agreement
- Other people's private messages, unless they've agreed
What to do instead: placeholders
You rarely need the real details for AI to help. Replace them with placeholders, get the draft, and add the real details yourself afterwards:
A customer, [Customer A], ordered [product] on [date] and it arrived damaged. Draft a reply offering a replacement.
The answer is just as useful, and nothing private left your computer.
Set the tool up before you use it for work
Five minutes in the settings is worth it:
- Find the data or privacy controls and decide whether your chats can be used to improve the model.
- Decide whether you want chat history kept. Some tools offer temporary chats that aren't saved to your history.
- Turn on two-factor authentication for the account.
- If several people use AI for work, consider a business plan with an admin console instead of personal accounts.
- Write down what you chose and when, so you can check it again when the tool changes.
Check before you use any answer
Privacy is half of safe use; accuracy is the other half. AI tools can produce confident, wrong answers: invented facts, wrong figures, links that don't exist. Before anything goes to a customer or into a decision, check names, numbers and dates against your records, open any links, and confirm legal, tax or safety points with an official source. Our five-minute check covers the routine.
Watch for AI-assisted scams
Scammers use the same tools. Messages that look like they come from a supplier, a colleague or your boss can be well written and convincing, and voice cloning makes phone calls harder to trust. Two rules prevent most losses:
- Never change bank details for a payment because of an email or message alone. Confirm by calling a number you already have.
- Agree a check for urgent money requests, such as a call-back or a code word, so "the boss" asking for an urgent transfer gets verified.
A one-page policy for your team
You don't need a long document. A useful AI policy for a small team answers five questions:
- Which tools are approved, and on which accounts?
- What must never be pasted in? (Use the list above.)
- What has to be checked by a person before it's sent or published?
- When do we tell customers that AI helped? (For example, if a reply was drafted by AI but checked by staff, many businesses don't mention it; if a chatbot answers directly, say so.)
- Who do people ask when they're unsure?
Get the checklist
The AI Safety & Privacy Checklist for Work puts all of this on a few pages you can print: the never-paste list with safe alternatives, a tool setup checklist and settings record, the checking routine, scam rules for payments and bank-detail changes, and a fill-in one-page AI policy for your team.
Setting up AI for a whole small business? The Complete Starter Bundle includes the checklist plus the 7-Day AI Workbook, 100 prompts, the brand voice kit and our beginners' guide for $39.
Helping a parent or older relative get started safely? AI for Beginners Over 50 ($9) covers privacy, checking answers and AI scams in large clear print, including a family code word for urgent money requests.
Not ready to buy? Start with the free guide, Your first week with AI: ten prompts and a privacy checklist, plus a five-day email course.